Welcome back for a new edition of Guru Gems. After putting out an article every Sunday for a year and a half (I have not missed a single one for 77 weeks straight!), today's post is a moment to look back, and also to look ahead.

On the menu for today:

• 📚 Guru Gems: The Best Of

• 💎 An honest look at the performance

• 👀 My watchlist from a different perspective (and why Visa stands out)

Let’s go!

📚 The Best of Guru Gems

There have been quite a few new subscribers over the past weeks/months (Thank you and Welcome! Guru Gems is closing in on 600 subscribers!), so I thought it would be good to share some of the ‘Best Of’ articles for anyone keen to learn from the best long-term investors.

3 most popular Guru deep-dives

1/ Norbert Lou - Why less is more, and how focus drives extraordinary returns. Also where the idea for the PayPal position started.

2/ Chris Hohn - On fortress businesses, and why S&P Global sits in two of the world’s most disciplined portfolios.

3/ Seth Klarman - On downside protection, patience, and the art of not blowing up

Physics for Investors

The ‘Physics for Investors’ series has also been quite popular. Here is the latest one, on value traps and why some cheap stocks stay cheap:

The quarterly Guru portfolio updates

And of course the posts at the core of Guru Gems, where I dig through the Gurus’ 13F filings to find the next Gem. Here is the most recent one:

And if you still want more…

The best way to find your way around all Guru Gems posts is the Archive, where everything is listed by category.

Before we continue: I would like to make Guru Gems even more useful for you, so I'd love to know which topics you would like to see more of. It only takes one click:

💎 An honest look at the performance

I started Guru Gems as a learning journey, trying to distill the frameworks and thinking process of the best long-term investors.

I have been putting my learnings to practice with two real-money portfolios.

As you can tell from the dashboard below, the performance of the Guru Gems portfolio is quite underwhelming.

It is lagging the S&P 500 by a wide margin this year, despite holding a few names that had a very strong run (Google +115%!)

There were certainly mistakes along the way, and I wrote about the biggest ones in my 'Podium of Errors' post (the gold medal went to my position sizing on Constellation Software).

But when I go through the portfolio name by name, I am still convinced about the companies I own. Here is why:

1/ Quality has been out of favor for over a year

The Guru Gems portfolio tilts towards ‘quality‘ companies: high returns on capital, strong balance sheets and steady growth. Over the past year, that was not where the returns were. The market has been all about momentum, especially in AI and semiconductors.

J.P. Morgan Asset Management’s latest Factor Views (August) describes the quality factor as “continuing its worst stretch since the post-COVID-19 speculative bubble and one of the worst in history”, while momentum “extended its best multi-year run since the dot-com bubble”.

Terry Smith described in his half-year letter a market “dominated by momentum rather than any fundamental factors like profitability, returns on capital and growth” , and added that momentum is now “at a 30 year high and more extreme than in late 1999 just before the Dotcom bubble burst.”

Fundsmith trailed the MSCI World by 14 percentage points in the first half of the year, and Smith turned over more than half of the portfolio, writing that he will “take more account of momentum” going forward.

2/ My laggards are cheaper than they have been in years

According to the same J.P. Morgan report, the quality factor is now “one standard deviation inexpensive globally vs. history, and more inexpensive than that within U.S. markets.”

I see the same thing in my own portfolio. S&P Global for example trades at about 20x forward earnings, compared to roughly 27x at the end of last year and 30x at the end of 2024.

Constellation Software is still around 30% below its 52-week high, even though it kept compounding adjusted EPS and free cash flow per share at roughly 20%.

3/ Long-term Gurus see the same gap

David Poppe at Giverny Capital Asset Management, who used that Constellation Software example in his Q2 letter (which I covered in August), sums it up like this:

“We’re generating a good return, but I think our companies overall are performing better than their stock prices would indicate.”

Some Gurus are also putting money to work on that gap. In Q2, while the market was de-rating S&P Global, Bill Ackman, Chris Bloomstran, Triple Frond Partners and John Armitage all opened new positions in S&P Global, and Pat Dorsey added to his.

4/ Warning signs, and why I’m holding 25% cash

Michael Burry has been warning about excessive valuations in AI names for many months with what he calls a ‘cornucopia of risks’ piling up on top of it.

Earlier this week he wrote that he is “moving timelines up” and now thinks the AI bubble “may burst sooner than later”. He is not the only one: over the past months I have shared similar concerns from Seth Klarman, Paul Tudor Jones, David Rolfe at Wedgewood and David Poppe.

That is also why I’m still holding 25% cash. It obviously has not helped the performance in a rising market, but it will come in handy when opportunities present themselves.

“However, always remember, my shorting and put buying - it is not for everyone. For many, holding stocks that are already cheap and raising cash for future opportunities is well and good enough.”

— Michael Burry, 28 September 2026

And the Magic Formula portfolio?

On a more positive note, my Magic Formula portfolio is still ahead of the S&P 500 this year (+13.2% vs +12.8% as of last Friday). The margin is small, but it got there without a single semiconductor or AI infrastructure name, in a year where semiconductors have grown to about a quarter of the S&P 500.

Subscribe now

→ If you enjoy this newsletter, please subscribe and like this post 🤍

→ If you find value here, consider buying me a virtual coffee ☕

👀 Buying 52-Week Highs: A Rule Breaker Look at My Watchlist

Earlier this week I was listening to the Behind the Balance Sheet podcast by (always great interviews!) where he interviewed David Gardner, co-founder of The Motley Fool and author of a very interesting book called ‘Rule Breaker Investing’, which I had also read a few months ago.

Here is the interview:

Buying the 52-week highs

One of the six traits of a Rule Breaker stock is ‘stellar past price appreciation’

“Most people hear those harmful 4 words ‘buy low, sell high’ and I’ve always said: buy high and try not to sell at all”

This may sound counterintuitive, but Gardner argues that stocks that win, usually keep on winning ("What do winners do? The answer is they win.").

“Let’s look at the 52 week highs in a world that’s often looking for the discarded cigar butts and the buy lows out there”

So in that spirit, I took the current Guru Gems watchlist and, instead of sorting it by the names that are down the most from their 52-week high, I looked at the ones that are closest to it.

Here are the 7 names on the watchlist that are within 10% of their 52-week high:

Why Visa stands out

Gardner makes it clear in his book that any one of the traits - including the stellar past price appreciation - does not stand on its own.

The key to Rule Breaker investing is to “look for the Rule Breaker traits together, working in concert”. And this is where Visa stands out.

How Visa scores on Gardner’s six Rule Breaker traits

• Top dog and first mover in an important, emerging industry: Visa is the largest international payment network in the world and processes nearly twice as many transactions as Mastercard. Payments is not exactly an emerging industry, but digital payments only overtook cash globally a few years ago, and Visa is now building out stablecoin and agentic commerce offerings. Half a point.

• Sustainable competitive advantage: the more consumers carry a Visa card, the more merchants need to accept it, and the more merchants accept it, the more consumers want one. I used Visa as the textbook example of a feedback loop in Physics for Investors #2. ✅

• Stellar past price appreciation: the stock is up about 22% since its 52-week low of ~$294 this spring, and within 7% of its 52-week high. ✅

• Good management and smart backing: Gardner prefers founder-led companies, which Visa is not. In his book, smart backing refers to the angel investors and venture capitalists who fund young companies. Visa is long past that stage, but the closest equivalent today would be the Gurus who have made it one of their largest positions, and there are quite a few of them (see Guru endorsements below). Half a point.

• Strong consumer appeal: there is a good chance there is a Visa card in your (virtual) wallet right now. ✅

• Grossly overvalued according to the financial media: Morningstar puts its fair value at $330, below this week’s close of $360, and rates the stock as overvalued. Gardner’s answer to that: “thank you very much”. ✅

Guru endorsements

• François Rochon has owned Visa since 2010. In a WealthTrack interview last November, Rochon is asked to name the company in his portfolio that best represents “the best of the best”. His answer: “I would choose Visa.”

  He bought it after the stock fell 25% on fears of new regulation on Visa’s fees, paying around 16 to 18 times earnings, and by his estimate earnings have grown by more than 15% a year since. When he bought it, he remembered a line from Warren Buffett in John Train’s The Money Masters:

  “Warren Buffett said that the ideal business is the one that receives a royalty on the growth of others. And I thought Visa was a perfect example of that.”

• Bill Ackman opened a position in Q2 (about 5% of Pershing Square’s portfolio). In his Q2 letter he wrote: “In our view, Visa and Mastercard are among the highest-quality businesses in the world.” He also addressed the biggest worry around the stock: “We believe stablecoins represent an opportunity for the card networks rather than a threat.”

• Visa is the largest position for John Armitage (14%) and ValueAct (14%), and the second-largest for Chris Hohn (20%), as I covered in my Q2 13F update.

• Not everyone is buying though: Terry Smith cut his long-standing Visa position in Q2 to help fund a new position in Mastercard.

Valuation

On an EV/EBIT basis, Visa trades at about 25x, slightly below its average of 26.4x since 2018, and well below the 35x+ it reached in 2021.

In hindsight, the moment to buy was late March and early April, when the stock dropped to its 52-week low and the EV/EBIT multiple fell to around 22x. Ackman’s letter mentions exactly that de-rating (to “22 times next twelve months’ earnings”), and his 13F shows he opened the position during the second quarter.

I realize it's a bit ironic to be looking for a cheaper entry point in a section about buying 52-week highs :)

Key risks

• Regulation and litigation: Visa took $2.6 billion in litigation provisions in fiscal 2025, and Europe is pushing ahead with a digital euro, partly to reduce its reliance on Visa and Mastercard (together 61% of card payments in the eurozone).

• New payment rails: stablecoins and agentic commerce could route payments around the card networks. But as Ackman and others point out, the opposite may happen as Visa is positioning its network to handle the future of money movement.

📌 I’m starting a small position in Visa

Visa scores 5 out of 6 on Gardner’s traits, has some of the best Gurus behind it, and trades slightly below its long-term average valuation. I’m starting a small position in the Guru Gems portfolio, at around 2%.

At 25x EV/EBIT it is fairly priced rather than cheap, and the gold medal on my own Podium of Errors went to position sizing on Constellation Software. So I’m starting small, which leaves room (and cash) to add if the market gives me another chance like the one in late March and early April.

That’s it for this week’s edition. As always, thank you for following along my learning journey!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome to Round 12 of the Magic Formula Portfolio!

This is a series within Guru Gems where I test whether Joel Greenblatt’s Magic Formula still works.

The rule for the Magic Formula portfolio is straightforward: hold each name for a year, then sell it and buy whatever ranks highest on the screener that week.

This month the screener kept two of the three names whose year was up.

Quick reminder if you are new to Guru Gems: while studying investing masters, I am creating two real-money portfolios to put my learnings into practice.

To learn more about my Guru Gems journey, start here.

Magic Formula Refresh

In ‘The Little Book that (Still) Beats the Market’, Joel Greenblatt suggests that a regular investor can beat the market averages by buying a group of ‘good’ companies at ‘cheap’ prices.

• ‘Good’ → High return on capital

• ‘Cheap’ → High earnings yield

Since he expects the magic formula to work on average, Greenblatt suggests to own a basket of 20-30 stocks, to buy this basket over the course of a year, and to hold each stock for exactly one year.

So that’s exactly what I’m doing…

You can read more on Joel Greenblatt and his Magic Formula here.

You can also download my Joel Greenblatt e-book for free:

Joel Greenblatt e-book by Guru Gems

494KB ∙ PDF file

Download

Download

One Year and Two Months in: Current Magic Portfolio Performance

On a YTD basis, the Portfolio returned +13.3%, just ahead of the S&P 500 (+12.8%).

A few names took a serious beating this week, in part driven by the buzz around Meta’s Muse AI agent. All aggregator companies, digital marketplaces and names that rely on digital ad revenue were sold off in the days following the announcement.

Here is the latest picture of the portfolio, which you can also access at any time here.

“If you are able to stick with the magic formula strategy through good periods and bad, you will handily beat the market averages over time” — Joel Greenblatt

Subscribe now

Cohort 3: One Out, Two Staying

Greenblatt’s rule is to hold each stock for exactly one year, then sell.

This week I should sell the names I bought on 29 September 2025: Harmony Biosciences Holdings (HRMY), Abercrombie & Fitch (ANF), and H&R Block (HRB).

Two of those (HRMY, HRB) still rank very high on the current screener, so I’m keeping them in the portfolio (it’s as if I sold them and bought them again).

Abercrombie (ANF) is the only one that goes out this round, leaving with a 62% return, which is not too bad.

H&R Block: Cheaper Than When I Bought It

HRB (Market Cap $5.2B) is down about 16% since I bought it. Over the same year its return on capital rose from 39.4% to 42.9%, and its EV/EBIT fell from 9.0x to 7.3x.

The business improved, the price fell, and the stock now ranks higher on the formula than it did when I bought it.

I ran the latest Magic Formula list (top 30 at a $1B minimum market cap) through four rankings: return on capital (ROCE) with EV/EBIT, ROCE using the three year average, then both again using P/FCF instead of EV/EBIT. HRB came between third and fifth on all four.

A reminder of what HRB does: H&R Block has been preparing US tax returns since 1955, through retail offices, franchises and a DIY digital product. Assisted preparation is 65% of revenue, the operating margin is 23%, the five year average ROCE is 36%, and the share count falls about 7% a year.

The market prices HRB as if software and AI are about to take the assisted business apart. The FY26 results, published in August, do not show that happening yet: revenue grew 4.9% to $3.95B, net income 21% to $736M, the company held its share of the assisted market for a third year running, and it retired 7.9% of its shares.

Harmony Biosciences: Up 50% and Still Ranking

Magic Formula metrics: ROCE 22% (ROIC 34%); EV/EBIT 8.2x; P/FCF 6.8x; Net cash

HRMY (Market Cap $2.4B) has gone up 50% since I added it last year, and it is still cheap enough to rank between seventh and twelfth on the four lists I mention above.

Harmony sells WAKIX (pitolisant), a treatment for narcolepsy, and the recent Q2 was a very strong quarter:

The key risk for HRMY is that almost all of Harmony’s revenue is derived from a single product. Any regulatory setback, safety concern, or loss of market exclusivity for WAKIX would severely threaten the company's financial viability.

HRMY is trying to mitigate that risk by expanding the indications for WAKIX, and more importantly by looking for strategic acquisitions that could advance the pipeline, like the 2024 acquisition of Epygenix Therapeutics.

The New Magic Formula Stock

As I explain in detail in my first Magic Portfolio post, I use Joel Greenblatt’s Magic Formula screener combined with financial ratios data to identify the top stocks in terms of combined rank for Return on Capital (using ROCE) and Earnings Yield (using EV/EBIT).

I extend the analysis with additional criteria: P/FCF (an alternative valuation lens to EV/EBIT), Revenue 3yr CAGR (is the company still growing?) and Net Debt/EBITDA (how much debt is on the balance sheet?).

With two names staying, only Abercrombie's slot is open.

Bath & Body Works (BBWI, Market Cap $3.4B)

Magic Formula metrics: ROCE 31% (3yr: 34%); EV/EBIT 6.9x; P/FCF 3.3x; Net Debt/EBITDA 3.0x

Bath & Body Works sells soap, candles and body care through more than 1,900 stores, mostly in North America. Stores are 77% of sales, digital 19%, international only 4%.

It ranked fifth of thirty on the latest list, and second and third on my two cash flow rankings. At a P/FCF of 3.3x it is one of the three cheapest names on the screen.

Why it is cheap

Sales are shrinking. The company guides to a 4.0% to 2.5% decline this year, and its share of the North American bath and shower market has fallen to 10.3% from 14.2% in 2021 while Dove and Olay held theirs.

Morningstar rates it no moat for that reason, and net debt sits at 3.0x EBITDA.

This probably summarizes why you can buy BBWI at four and a half times earnings.

What the price is ignoring

New leadership is implementing a turnaround plan (“Consumer First Formula”) which shows early promising signs.

Second quarter sales fell 2.3% to $1.5 billion, but operating margin rose 370 basis points to 14.8%. Digital sales grew for the first time since 2021. Full year EPS guidance went up to $2.60-$2.80, helped by a $250M cost reduction program running ahead of plan.

Morningstar puts BBWI’s fair value at $54 against a share price around $16. Fair value estimates are obviously based on a lot of assumptions, but the gap tells you how much decline is already in the price.

One name in, one out, two staying, and the portfolio remains at 25 stocks.

That’s it for this week! Please like or share if you enjoy reading Guru Gems.

You can follow me on Substack @gurugems and X @guru_gems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

Last month I went through the Q2 13F filings and wrote about the trends across the Gurus I track. Alphabet is still the ultimate Guru Gem, the toll booths (Visa, Mastercard, Moody's, S&P Global) kept attracting buyers, and Netflix and Uber showed up in many portfolios.

In today’s post I will highlight 3 names which have stayed somewhat under the radar:

• 🏗️ Builders FirstSource (BLDR): a building materials company near its 52-week low, with two value investors buying

• 🛒 Sea Limited (SE): the largest holding in a Davis Advisors fund, and a company Howard Marks bought again

• 💉 Becton Dickinson (BDX): Marc Werres’s first new position since 2022, and the only ‘special situation’ in the Hinde Group portfolio

Let’s jump in.

New to Guru Gems? Start here.

🏗️ Builders FirstSource (BLDR)

Builders FirstSource is the largest supplier of building materials to professional homebuilders in the United States. It manufactures roof and floor trusses, wall panels, windows, doors and millwork, and it distributes lumber and everything else that goes into a house, from approximately 570 locations across 43 states.

The numbers are ugly

Second quarter net sales fell 9% to $3.9 billion, gross margin dropped 260 basis points to 28.1%, and adjusted EBITDA came down 34.9% to $329 million. The company posted a small net loss of $3.9 million against net income of $185 million a year earlier.

Management cut full-year guidance to $14.0-$14.8 billion in sales and $1.0-$1.2 billion in adjusted EBITDA, assuming single-family starts down nearly 7%. [‘Housing starts’ is an economic indicator that reflects the number of privately owned new houses on which construction has been started]

Net debt to EBITDA has gone from 2.3x a year ago to 3.6x, as the result of EBITDA collapsing under the same amount of debt.

The stock closed at $58.12 on 18 September, roughly 2.8% above its 52-week low of $56.56.

Guru endorsements

Christopher Bloomstran (Semper Augustus) initiated a position in BLDR in Q1 2026 and then increased it by 143% in Q2. It is now his seventh-largest holding at 5.6%, just behind Dollar General, Deckers Outdoor and Dollar Tree. Bloomstran is known for writing a hundred-page annual letter dissecting Berkshire’s book value, and Berkshire remains his largest position at 25.6%.

Peter Keefe (Rockbridge Asset Management) increased his BLDR position by 10% in Q2, to a little over 2% of his portfolio. Across roughly 30 holdings, Builders FirstSource and Copart were the only two existing stock positions he added to.

I covered Keefe and his Copart thesis in issue 17, and Copart came up again in last week’s edition.

François Rochon, Miller Value Partners and Tweedy Browne also appear, but at very small sizes of the portfolio.

One more piece of context that I’ll come back to later. Edgar Wachenheim’s Greenhaven Associates has roughly 38% of its $8.7 billion portfolio in four homebuilders: Lennar, Toll Brothers, PulteGroup and DR Horton. BLDR is not in the list, though he is clearly making the same directional bet on US housing.

Why it might work

The bull case for BLDR is based on the shortage of housing in the US. Homes in the US have been underbuilt by one to four million since 2010, according to BLDR’s own estimates.

BLDR has also used the downturn to shift its mix towards higher-margin value-added products like prefabricated components and millwork, now around 48% of sales.

Since August 2021 the company has also repurchased nearly 50% of its shares outstanding.

How much did the Gurus pay?

Bloomstran and Keefe were both buying through the second quarter, which BLDR ended at $89.48. At $58.12 today, anyone following them is paying 12% less than the Q2 low.

On an EV/EBIT basis, BLDR doesn’t look very cheap compared to its historical valuation:

📌 BLDR is a pass for me

Revenue has fallen every year since 2022 and is guided lower again in 2026, net leverage has climbed to 3.6x against 2.3x a year ago, quarterly free cash flow is down to $32 million, and the housing outlook got worse rather than better between the first and second quarter guidance. Add to that an industry I know very little about and the result is that I will skip it for now.

Subscribe now

→ If you enjoy this newsletter, please subscribe and help others discover it by clicking 🤍

→ If you find value here, consider buying me a virtual coffee ☕

🛒 Sea Limited (SE)

Sea Limited is a Singapore-based company running three businesses across Southeast Asia, Taiwan and Brazil:

• Garena, a gaming studio, built around Free Fire

• Shopee, the region’s leading e-commerce marketplace by gross merchandise value

• Monee, its digital financial services arm, mostly consumer and small-business lending

A very good quarter

Second quarter revenue was $7.8 billion, up 48% year on year, with net income of $458 million and adjusted EBITDA of $917 million, up 11%.

Shopee’s gross merchandise value (GMV) grew 28% to $38.3 billion, its eighth consecutive quarter of sequential growth, and management reaffirmed full-year GMV growth of around 25%. Core marketplace revenue (fees and advertising) grew 66%. The VIP programme passed 15 million members, up 45% in a single quarter, and those members already account for 24% of Asia GMV with about 80% monthly retention.

Monee’s loan book reached $11.1 billion, up 62%, with 90-day non-performing loans at 1.0%, though margins are compressing as it pushes into new products and countries.

Garena is the softer piece. Bookings, meaning the cash players actually spent inside its games, still grew 15% year on year, but the business still leans heavily on one title. Free Fire is in its ninth year, and Garena has yet to prove it can launch another game at that scale.

Despite the strong quarter, the stock is at ~$102, which is around 48% below its 52-week high.

Guru endorsements

Davis Advisors

Chris Davis holds SE at 0.83% of the firm-wide 13F, which may look trivial. However, inside the Davis Select International ETF (DINT), a 28-stock fund run by Danton Goei that looks for high-conviction, best-of-breed businesses outside the US, Sea Limited is the single largest holding at nearly 6% as of 15 September.

“[The companies we hold] have solid earnings, yet are 13% less expensive than the benchmark. This valuation gap is among the widest we have ever seen.”

— Danton Goei, Davis Select International ETF

I profiled Chris Davis and the Davis approach back in issue 7:

Other Gurus

Howard Marks’ Oaktree first bought SE in Q2 2025, then added higher, then sold everything, and in Q2 came back to the stock again, buying it at roughly half the price but 4 times the previous position. The Q2 2026 position is 1.61% of the reported portfolio.

Also worth noting: Tiger Global still holds SE at 5% of its portfolio, the largest weighting of any filer on the list, even after reducing by 18.5%. Viking Global added 19.7%, and Stephen Mandel’s Lone Pine opened a new position.

How much did the Gurus pay?

The stock closed this week 6% above the Q2 high, but as mentioned earlier still significantly below its 52 wk high and also well below Morningstar’s fair value estimate of $152.

📌 Sea is already on my watchlist

I will move it up the list. It’s a genuinely interesting business and I’ll be taking a closer look before taking a starter position.

💉 Becton Dickinson (BDX)

And now one I’ve been watching for a while.

Becton Dickinson makes the unglamorous consumables of modern medicine: syringes, catheters, infusion pumps, prefillable syringes for injectable drugs, surgical products. Roughly 90% of hospital patients encounter at least one of the more than 45 billion products BD manufactures each year. More than 60% of revenue comes from the US.

It is a business most investors find boring, and it has been a poor stock for five years.

Marc Werres bought it as a special situation

In May I wrote a deep-dive on Marc Werres, founder of Hinde Group and one of the best investors almost nobody has heard of. Since inception in August 2015 through the end of Q2 2026, Hinde has compounded at 20.4% net per year against 14.3% for the S&P 500, with fewer than ten positions and no leverage.

Werres divides his universe into two buckets. Compounders are great businesses he intends to hold indefinitely, and they make up most of the portfolio. Special situations are less-than-great businesses trading at a discount, where a specific catalyst is expected to close the gap within about two years.

In July 2025 he opened a position in BDX. It was his first new position since 2022, and the only special situation in his portfolio.

Here is how he described the business in his Q3 2025 letter:

“In many of its product lines, BD has leading, if not monopoly-like, market shares that have been stable for decades... BD’s market power is plainly evident in the levels of its gross margin (54.7%), operating margin (25.0%) and return on tangible invested capital (40%+).”

“BD falls just short of a great business... All that is to say while BDX is a special situation investment, it is an especially attractive special situation because BD is almost a great business.”

The catalyst has already happened

The specific event Werres was underwriting was the Reverse Morris Trust transaction with Waters Corporation.

BD spun off its Biosciences and Diagnostics Solutions business, about $3.3 billion of revenue, and simultaneously merged it into Waters. The deal closed on 9 February 2026. BD received $4 billion in cash, and shareholders received 0.135 Waters shares per BDX share, worth $44.86 per BDX share at Waters’ opening price that day.

Werres added to the position in Q4 2025 and sold the Waters shares he received. BDX contributed 1.95% to Hinde’s gross return in 2025 and another 0.84% in Q1 2026.

In his Q1 2026 letter, with the stock around $160, he shared the following:

“BDX would have to trade at more than $230 per share today to bring the prospective return on the stock closer to a market-level return.”

He also noted that BD’s highest priority use of free cash flow right now is buying back stock, and that the remaining catalyst is the company consistently hitting its mid-single-digit organic growth target, which he expects from fiscal 2028.

There was no update on BDX in the Q2 2026 letter, which I assume means that the position remains unchanged. (Hinde does not file a 13F, so the partner letters are the only insights we get.)

Guru endorsements

Werres is not alone, and the largest holder on this list is an interesting one.

Greenhaven Associates opened a brand new BDX position in Q2, making it the fourth-largest holding at 10% of an $8.7 billion portfolio.

Greenhaven is run by Edgar Wachenheim III, author of Common Stocks and Common Sense. Portfolios he managed returned an average of more than 18% annually from 1990 to 2014, with conservative stocks and no leverage, and his top ten positions are 79% of the portfolio. He looks for cyclical, cash-generative businesses trading well below normalised earnings, and he is explicit about being early: he was buying housing in 2011 and airlines in 2010-2012.

This is the same investor with 38% of his portfolio in homebuilders, which is where I started this post. Wachenheim is placing two large, unfashionable bets at the same time.

First Pacific Advisors added 6.6%, First Eagle added 15.8%, and Bill Nygren at Oakmark added 5.8% to a small position.

Where things stand now

The third quarter, reported on 6 August, was the first clean quarter for the slimmed-down company:

More than 90% of the portfolio delivered high-single-digit revenue growth, and management raised the midpoint of full-year EPS guidance to $12.62-$12.72. They have returned $3.1 billion to shareholders so far this year, $2.3 billion of it in buybacks, taking the diluted share count from 287 million to 275 million.

Morningstar rates BD a narrow moat with a $225 fair value estimate against the $180.94 close on 18 September, and wrote after the quarter that “the results should pave the way for some rerating of the share price.”

Together with Werres’s $230 marker, that puts two independent estimates roughly 24% to 27% above today’s price.

How much did the Gurus pay?

Greenhaven built its entire $860 million position inside the second quarter, and Werres has been holding since July 2025.

📌 BDX goes on the watchlist

I already own Baxter, added in September 2025 after my Pzena deep-dive, and that is another medtech turnaround where the thesis hasn't fully played out yet. I'd rather not hold two positions in the same corner of the same sector before I've seen how the first one resolves.

That’s it for this week’s edition. As always, thank you for following along my learning journey!

You can follow me on Substack @gurugems and X @guru_gems for more insights.

Until next week!

Share Guru Gems

P.S. Shoutout to who is building out ownershiptrace.com - Check it out!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back and thank you to all new subscribers for joining Guru Gems!

Today I’m sharing my thoughts on 2 things that caught my attention this week: a long interview on one of my favorite podcasts, and an earnings update of one of my Guru Gems.

2 minute AI audio version:

New to Guru Gems? Start here.

1️⃣ Chris Begg: “Alphabet remains one of the most extraordinary businesses ever built”

As you may know, I’m a big fan of the Richer Wiser Happier podcast series. This week I listened to a recent episode with Chris Begg, who runs East Coast Asset Management and teaches the Security Analysis course at Columbia Business School (the same course that Ben Graham taught to Warren Buffett in 1951).

Long-time Guru Gems readers may recognise this name. Back in April 2025, in the second edition of this newsletter, I quoted Begg from an earlier episode of the same podcast (RWH #56):

“Volatility is our friend, always. The price we pay will determine the rate of return.”

In that issue I listed the Gurus who owned Alphabet, and East Coast had 15% of the portfolio in it. Today it is 17% and still Begg’s largest position. Alphabet is also the largest holding in the Guru Gems portfolio, so I listened very carefully to this episode.

East Coast Asset Management portfolio

Before we get into the interview, let’s have a look at Begg’s latest portfolio. Looking at the Q2 13F filings, it shows 65 positions in total.

What’s a bit strange is that in the interview Begg says “we own less than ten companies in our portfolio.” There is a long tail of companies and ETFs which are less than 0.2% of the portfolio, so it could be that many of these are either legacy positions or companies they recently started looking into.

The seven names listed above are about two thirds of the reported portfolio, and he mentions nearly all of them in the interview. Begg calls them his “Grove of Titans”.

Constellation Software, a recent position and also one of my Guru Gems, trades in Toronto so it doesn’t show up in the 13F filings.

Pillars, clouds, and graphs

Begg wants three pillars in every Grove of Titans business: 1) a moat that is widening, 2) secular tailwinds driving the top line, and 3) a human element, meaning operators with a history of intelligent capital allocation.

The best time to buy these titans, according to Begg, is when there are ‘clouds’ around the business.

“When you get the quality at a reasonable price, there’s usually clouds.”

A cloud, as he puts it, is “a perception that something might change versus evidence that it is” and his research is organised around one question:

“Much of our research effort is devoted to a simple question. Are the clouds temporary or permanent?”

Alphabet, the ultimate ‘graph’

Begg describes Alphabet as a graph. A graph is a mathematical structure used to model pairwise relations between objects (read more here).

“At its core, it is a graph designed to organize the world's information and make it universally accessible and useful.

[…] Nodes and edges populating across a domain, [producing] increasing returns to scale […] instead of the diminishing returns most businesses run into as they get bigger.”

He argues that this is what a lot of people misread about the last decade of technology: the big platforms did not slow down as they scaled, their growth accelerated.

The two clouds for Alphabet were the threat to search and blue links as well as the antitrust case, and East Coast owned the business at fifteen times earnings while both were still hanging there. Note that this is also when Guru Gems bought Alphabet… ;)

Search volumes went up with AI rather than down, the ruling went their way, and the stock rerated. In his words: "it was almost like we woke up one day and the cloud was gone."

Begg still owns Alphabet as his largest position, and he believes Alphabet remains one of the most extraordinary businesses ever built.

He sees Alphabet as a collection of fantastic businesses (Google Cloud, YouTube, Waymo, …) with at the core a leading AI effort under Demis Hassabis at DeepMind that runs well past building LLMs, into work like AlphaFold.

"There’s a large AI effort that Demis has led, which is leading to all kinds of different applications we saw with AlphaFold, health care breakthroughs, AGI, ASI.

So we think there’s a lot of things that are misunderstood about their AI advantages and how broad it is.”

Constellation Software: ten years of homework

Begg has been studying Constellation Software for over ten years. He knows Mark Leonard and has had him in the Columbia classroom three times. He knew the business very well, but in all those years he never owned any shares, because it was always fully valued.

Then came the software apocalypse, what Begg calls “the biggest cloud of 2026”, and the stock fell 50% from its highs.

That is when East Coast went to work, studying the business even deeper and taking the whole team to Constellation’s investor day in Toronto to meet Mark Miller and the division heads. Then they bought.

“The cloud that is still resident around software is a perception that these assets could be disrupted versus any evidence that they are.”

Norbert Lou’s playbook

Listening to Begg talk about his Constellation purchase made me think of a similar story from a different Guru.

Last October I wrote about Norbert Lou of Punch Card Capital, who watched TGS-Nopec for seven years without buying a single share. Then Deepwater Horizon blew out in April 2010, Washington shut down drilling in the Gulf, and a company that earned roughly half its money there fell to six times after-tax earnings.

TGS-Nopec, in Lou’s words, “was priced as if the Gulf of Mexico would never re-open.” He started buying in June, and the stock rose 90% over the back half of that year.

“When you can be really patient and selective, you have the luxury of waiting for that confluence of multiple forces.” — Norbert Lou

Constellation’s risk to watch

Begg says that what he is watching now is churn at Constellation’s individual businesses. If that starts accelerating, the cloud is real and he is wrong.

But so far he is finding the opposite, with AI something that can be layered on top of these businesses rather than the thing that kills them.

Tesla and SpaceX: the Value 3.0 lens

Tesla is a stock most value investors cannot picture in their portfolio.

In my 10 May issue I wrote about Christopher Tsai and his Value Investing 4.0 essay. Tsai has Tesla as his largest position (22%) and believes it is deeply misunderstood. Begg is now in the same camp, at nearly 13% of his portfolio after adding again in Q2.

I think the interesting part to understand is not whether he is right about Tesla and SpaceX, but how someone who teaches Ben Graham’s course ends up here at all.

The framing he uses is ‘Value 3.0’, a label he shares with James Anderson. Graham was 1.0, a dollar of assets for fifty cents, margin of safety in the balance sheet. Buffett and Munger were 2.0, a wonderful business at a fair price, margin of safety in the discounted cash flows.

In 3.0 the margin of safety comes from the end state. Begg suggests Nick Sleep was already doing this with Amazon, which looked expensive on every 1.0 and 2.0 measure and turned out, in hindsight, to be one of the cheapest companies in the world.

He says that the one thing that does not change across the three is asymmetry:

“Everything that a value investor does and thinks about has to be asymmetric,”

So rather than modelling next year he asks what a business is worth in ten and works backwards, which is close to what Bill Miller once told William Green: he was not buying at a discount to what things are worth, but at a discount to what he believed they would be worth.

It also explains why he has no regrets about passing on Tesla in the early years. Back then you were still underwriting whether the thing could be built at all. He waited until the platform existed and only then started counting what could be built on top of it. He sees five businesses inside Tesla, five inside SpaceX, and he expects a merger of the two within twelve months.

I don’t think I will be adding either name to my watch list any time soon, but I do want to keep an open mind about the frameworks used by these investors.

→ If you enjoy this newsletter, please subscribe and help others discover it by clicking 🤍

→ If you find value here, consider buying me a virtual coffee ☕

Subscribe now

My most recent Guru deep-dive is on Seth Klarman, one of the great quality investors. I recently updated the post with Q2 portfolio data and it is quickly becoming one of my most read posts, so check it out if you haven’t yet!

2️⃣ Copart buys ACV Auctions

Copart reported Q4 fiscal 2026 earnings on Thursday. It is one of my Guru Gems, and I first looked at it in August 2025 when I profiled Peter Keefe of Rockbridge Capital.

Last week I flagged three things I would watch in Copart's Q4 earnings.

The quarter itself

Fiscal Q4 was not exactly spectacular, with EPS of $0.35 against $0.41 a year ago and full year EPS of $1.55 versus $1.59. Full year revenue came in at $4.7 billion, a 0.4% increase over prior fiscal year.

The biggest news was the ACV Auctions acquisition, which I will cover in some more detail below.

First, a short update on the three points I was watching before the call:

1. Reallocating capital. Copart announced an all-cash deal to acquire ACV Auctions, a digital dealer-to-dealer marketplace selling over 800,000 vehicles a year, and plans to use its 275+ physical locations as staging and marshaling hubs for ACV’s land-light model.

2. Volume recovery. Global unit sales fell 2.9% in the quarter, with the US down 5.7% and international up 10%. Excluding a single lost customer account, Adair said US insurance assignments would have been up 2.3%. Total loss frequency reached 23.3%, a record for a second calendar quarter.

3. The balance sheet. Copart closed FY2026 with $4.5 billion in cash and held-to-maturity securities. Even after funding the ACV deal, CFO Leah Stearns said it retains more than $2 billion in cash plus a $1.25 billion undrawn credit line and no debt outstanding.

What Copart actually bought

ACV is a digital marketplace for wholesale vehicles, mostly dealer to dealer: $10 billion of annual gross merchandise value, 22,000 unique buyers, 70,000 transactions a month, plus a transportation arm, a floorplan lender (ACV Capital) and an inspection stack including VIPER.

Copart is paying $10.50 a share, valuing ACV at $1.9 billion, a 45% premium to where ACV traded on 10 August, the last day before media reports of a possible deal.

“ACV sells more than 800,000 vehicles each year and, importantly, operates with virtually no land of its own” — Jay Adair

Why they bought it

1. Totalled but not wrecked

In the earnings call, Adair talks about how complicated cars have become, with an interesting comparison: a military drone runs on about 3.5 million lines of code, an Airbus on 30 million, Windows 10 on 50 million, and a new Tesla on roughly 100 million.

“When we think about cars, they really are becoming computers on wheels. We believe total loss frequency will continue to go up.”

And it is indeed already going up. Total loss frequency hit 23.3%, a record for a second quarter, and average collision severity passed $6,300 a claim, up 8.8%.

The consequence, in Adair’s words, is that:

“More often than not now, you’re seeing cars that don’t look like they should have been totaled, but they’re economic totals. While they’re still drivable, while they’re repairable, they’re economically totaled.”

Copart’s buyer base was built for wrecks: dismantlers, rebuilders and exporters. The natural buyer for a drivable, lightly damaged car is a franchise dealer, and those are exactly the buyers ACV will bring them.

Adair again:

“Think about every car that is not damaged or lightly damaged is going to be put in front of thousands of dealers.”

2. Growth without land

Copart has always grown by buying and building yards, spending $569 million on property and equipment in FY2025 and $337 million in FY2026. ACV adds more than 800,000 units a year without an acre.

Adair wants to run ACV's cars through Copart's existing yards, and it works the other way too.

When asked how a physical network helps a digital business, he described a dealer with 50 cars that nobody can efficiently collect:

"With Copart, we can move those vehicles over to Copart, and then a nine-car coming through can pick those up on their time and bring them down to Mexico or bring them to a port."

ACV's lower-end trade-ins are vehicles "our international buyers, especially Mexico, just love."

3. Buying what they could not build

This is not Copart's first go at whole car, meaning undamaged vehicles rather than salvage.

It launched Dealer Services back in 2007, and sellers other than insurers now make up about a quarter of its volume. But its own Copart Direct programme saw volume fall 34.5% this year, and the franchise dealers never really came.

Asked why the earlier attempts had not made a bigger splash, Adair replied:

"To get into those franchise dealers and get into the higher-end trades, I think does take a different product. ACV is a different product than Copart."

The deck says the same thing more politely: “ACV's […] dealer base adds upstream supply Copart does not currently address".

What it costs

So what does $1.9 billion buy them? ACV guided its own 2026 to roughly $850 million of revenue, $73 to $77 million of adjusted EBITDA and a GAAP net loss of $44 to $49 million. Copart has paid a 45% premium for a business that does not yet earn a GAAP profit, in a year when its own EPS went backwards.

On the earnings impact, Copart expects the deal to be neutral on EPS in the first full year and accretive from fiscal 2028. There are no numbers yet on near-term cost and synergies, but Adair is relying on their track record to create value:

“Copart has a very strong track record of driving strong return on invested capital across the businesses it has acquired.”

What’s next

Both boards have approved the acquisition and management expects to close by calendar year end.

When asked whether this rules out further M&A, Adair said:

“I don’t think this prohibits us from doing any future acquisitions. We’re looking at other businesses that we may want to acquire in the auction space.”

That’s it for this week! Please like or share if you find these insights valuable.

You can follow me on Substack @gurugems and X @guru_gems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

This is the third issue of Physics for Investors, a Guru Gems series where I revive my old passion for physics and apply a single idea from that world to how we think about businesses and markets.

In Issue 1, I used the second law of thermodynamics to explain why most businesses decay toward the cost of capital. In Issue 2, feedback loops explained why a rare few businesses escape that decay and compound instead.

Today’s issue is about a third state every value investor knows well: the stock that is neither decaying nor compounding. A stock that is simply stuck.

I will try to answer two questions about these stocks.

First, what does it mean for a stock to be stuck, and how does a stock that is truly stuck (holding real value the market refuses to recognize) differ from one that only looks stuck?

Second, if nobody can predict when any single stock comes unstuck, how should we invest in them at all?

And of course there will be a Guru. Today I introduce you to none other than Ludwig Boltzmann, the man who bridged the gap between the invisible world of atoms and observable macroscopic properties like temperature and pressure.

“I adhered scrupulously to the precept of that brilliant theoretical physicist L. Boltzmann, according to whom matters of elegance ought to be left to the tailor and to the cobbler.”

— Albert Einstein in the preface of his book ‘Relativity: The Special and the General Theory’

2 minute AI audio version of today’s post:

New to Guru Gems? Start here

🧊 The water that refuses to freeze

Everyone knows water freezes at 0°C (32°F). Except that in reality, it often doesn’t.

High up in the atmosphere, clouds are full of liquid water droplets at minus 10, minus 20 degrees Celsius. Very pure water can stay liquid down to about minus 40°C (which, coincidentally, is the one temperature where Celsius and Fahrenheit agree).

You can try a version of this at home: leave a smooth bottle of purified water undisturbed in the freezer for a couple of hours, take it out gently, and it may still be liquid several degrees below its freezing point. Then give the bottle a sharp knock, and ice crystals will shoot through it in seconds.

Physicists call this state supercooled, and it is the classic example of a metastable state.

Here is a ‘cool’ video explaining the physics of supercooling:

So what is going on? Below 0°C, ice is the more stable arrangement, so the water “wants” to freeze. But the freezing has to start somewhere; a first tiny cluster of molecules must lock into a crystal seed before the rest can follow.

Forming that first seed briefly costs energy before it releases energy. There is a ‘barrier’ to get over, and physicists call the height of that barrier the activation energy.

So the supercooled water is not in its most stable state, and it is not going anywhere either. It can stay like that for hours, sometimes indefinitely, until something kicks it over the edge: a knock, a speck of dust, an ice crystal dropped in. That something is called a nucleation event.

And when the transition finally comes, it can go fast. The bottle sits there doing nothing for hours, and then turns to slush in a few seconds..

Are you still following? :)

Let’s take a break from physics, and just keep these two ideas in mind: a barrier, and a trigger.

Now let’s talk about cheap stocks.

🪤 Supercooled stocks and value traps

Imagine a stock trading at $20 that you are convinced is worth $40.

If the stock is really worth $40, why is it still trading at $20? Why hasn’t the market corrected it already?

Metastability may suggest an answer: because there is a barrier. The pessimism is priced in, the sellers have already sold, everyone has accepted the narrative (“secular decline”, “uninvestable”, …), and nobody wants to be first. The market will not change its mind spontaneously. Something has to happen first.

Every value investor knows what this ‘something’ can look like: an earnings report that proves the pessimism wrong, or a new CEO, a product launch, a divestiture, an activist arriving, a cycle turning, a takeover bid. These are the market’s nucleation events.

Interestingly, the word used in finance for these events is the same word used in science: catalyst. In chemistry, a catalyst is defined as something that lowers the activation barrier without being consumed.

Is the stock stuck, or is it broken?

Before going further, I should point out where I think the physics analogy stops working.

When supercooled water gets its knock, it always freezes: thermodynamics defines which state is stable, so the destination is guaranteed.

A cheap stock doesn’t have such guarantee. While you sit there waiting for something to happen, the business itself can deteriorate. The share price stays at $20, but the $40 you thought it was worth erodes to $30, and then to $20.

So for every statistically cheap stock, there is really one question to answer: is it stuck, or is it broken?

A metastable bargain is stuck. The business is temporarily impaired, but its long-term economics remain intact. The normalized earnings power is still there. Something specific stands between today’s depressed state and a healthier future state, and that something can happen.

A value trap is broken. The business is deteriorating structurally. The ‘normalized’ earnings you are anchoring to no longer exist and there is no barrier to cross. It is just a business heading toward a permanently lower equilibrium. This is the decay I covered in Issue 1, now disguised as a stock that is stuck, and the disguise is why value traps fool so many investors: the price stands still while the value beneath it erodes, until one day the “cheapness” is gone without the price ever having moved.

Now let’s introduce our Guru.

Subscribe now

If you enjoy this newsletter, please help others discover it by clicking 🤍

🧔🏼‍♂️ The Guru: Ludwig Boltzmann (1844–1906)

Boltzmann was born in Vienna in 1844 and spent his career on one big idea: that the visible properties of matter (temperature, pressure, entropy) are statistics over countless invisible molecules.

There is a nice connection with the 2 previous Gurus.

Issue 1's guru, James Clerk Maxwell, laid the foundations of the kinetic theory of gases, and Boltzmann picked up that work and carried it much further.

Issue 2 featured Einstein's 1905 Brownian motion paper, which Norbert Wiener later turned into mathematics. That same 1905 paper would settle a thirty-year debate about whether atoms exist at all, and Boltzmann stood at the center of that debate.

Building on Maxwell’s work, Boltzmann derived what we now call the Maxwell-Boltzmann distribution: at any given temperature, the energies of molecules follow a predictable curve. Most molecules are ordinary. A few, out in the tail of the distribution, are energetic enough to do something rare: escape a liquid, trigger a reaction, or clear a barrier.

The interesting part about the distribution is that you can never say which molecule will clear the barrier, but you can always say, with great precision, what fraction will.

In 1877, Boltzmann also gave a statistical meaning to entropy, a concept I touched on in Issue 1: disorder is simply the number of ways a state can be arranged. The formula, S = k log W, is carved on his gravestone in Vienna.

For decades, Boltzmann defended the reality of atoms against the scientific establishment of his day, led by Ernst Mach and Wilhelm Ostwald, who considered atoms an unscientific fiction since nobody had ever seen one.

This lifelong scientific struggle deeply impacted his career and personal life. Boltzmann, increasingly isolated and in declining health, passed away on 5 September 1906, almost exactly 120 years ago this week.

Within three years, Jean Perrin’s experiments confirmed Einstein’s Brownian-motion predictions, and even Ostwald conceded that atoms were real.

Boltzmann held his unpopular position for thirty years and never got to see the proof arrive. A feeling that may be familiar to value investors :)

🎲 Betting on the distribution, not the molecule

Back to our cheap stock.

Owning a stock that you believe is cheap means you are making two bets:

1. You have to be right about the value: the stock is stuck, not broken.

2. You have to be right, or at least lucky, about the timing: the push has to arrive within your investment horizon.

Owning twenty-five of them changes the problem.

You are no longer asking “when will this stock re-rate?”, which nobody can answer. You are asking “how many of these 25 will cross their barriers this year?”

That second question does have an answer, and it is the exact question Boltzmann answered for molecules.

This is where Joel Greenblatt’s Magic Formula comes into the picture.

A basket of 20 to 30 good-and-cheap stocks, each held for exactly one year, with no exceptions. Some fraction will get their earnings surprise, their takeover bid, or their cycle turn within any given year.

Then you basically collect whatever fraction of the basket has ‘nucleated’, recycle the rest, and restock with freshly supercooled names.

There is even some academic evidence that fits this picture.

In a well-known 1997 study, Good News for Value Stocks, La Porta, Lakonishok, Shleifer and Vishny found that a significant portion of value stocks' outperformance is delivered in the short windows around earnings announcements, as value stocks keep delivering positive surprises for years after they were identified as cheap.

In other words: most nucleation events are boring quarterly reports, impossible to time for any single stock, but still happening enough across the population to have a meaningful impact.

Watching the barriers break

A bit over one year into my own Magic Formula Portfolio, I have been able to watch this process live:

• EverQuote (EVER) jumped 64% in May on a strong first quarter.

• Abercrombie & Fitch (ANF) did the same thing three months later, reporting earnings in August that beat expectations by a wide margin and raising its full-year guidance. The stock rose more than 30% in a matter of days.

• Catalyst Pharmaceuticals (CPRX) was acquired by Angelini Pharma at $31.50, a 58% return in my portfolio, delivered by a trigger that probably only a few saw coming.

• Rimini Street (RMNI) jumped about 30% on a legal settlement, but then gave most of it back. Clearly, not every knock is hard enough, and some pushes fail to clear the barrier.

I could not have picked these winners in advance, but I also did not need to. 21 of 25 positions are currently in the green, and the basket is up 22% YTD, despite individual disasters like YELP and LULU.

Speaking of Lululemon (LULU), they just reported a very poor quarter, cut the guidance once again, and the stock fell around 17%. It now trades roughly 80% (!) below its Dec 2023 high. Plenty of commentators call it the definitive value trap.

Others point out that a new CEO takes over this month, and that a quarter like this is what you would expect a company to report right before a new boss walks in.

Stuck or broken? I don’t have the answer. On the Magic Formula criteria, it still looks attractive (ROCE 32%; EV/EBIT 7.1). But it does start showing signs of a value trap (declining revenue and falling comparable sales).

The good thing is that with 25 equally-weighted names, the impact on my portfolio is limited. Lululemon still has 8 months in the portfolio, and when its year is up it gets sold regardless of what I conclude.

“If you are able to stick with the magic formula strategy through good periods and bad, you will handily beat the market averages over time” — Joel Greenblatt

Patience is still key

Greenblatt made it clear in his book that the formula goes through long stretches of not working, which is exactly when many investors give up, and so miss the payoff that comes later.

Some years simply produce fewer crossings than others. The stored energy does not go anywhere when that happens. It waits, and it accumulates.

That is worth holding in mind when you look at where value sits today. As I wrote in my Pzena Halftime post, the valuation gap between the cheapest and the most expensive parts of the market is about as wide as it has ever been.

To stick with the physics analogy: there are a lot of supercooled bottles in the freezer right now, and each one is holding more stored energy than usual.

💎 Deep value: curating the distribution

Since nobody can win the timing bet, the value bet is the one worth working on. And there is a spectrum of how much work you can put in.

On one end of the spectrum are simple filters, applied to the whole population.

Since the two failure modes of a cheap stock are the broken business and the unclearable barrier, I try to keep the most likely value traps out of my Magic Formula basket: companies whose revenues or earnings decline year after year (the classic broken profile, hence the Revenue 3yr CAGR and EPS 3yr CAGR checks I’ve added), and companies carrying heavy debt (hence the Net Debt/EBITDA filter).

At the other end sit the Deep Value investors, like Rich Pzena.

They build the entire investment process around one question: is there a barrier at all?

In a recent interview, Pzena explained how his firm screens on what they call price-to-normal earnings:

“Normal earnings we can define as what should the company earn over a full economic cycle... What’s at the top of the list are companies that are selling for a low price relative to what their history suggests they should be earning, and they’re generally not earning that. So you have a very rich universe to go in and analyze it and say, is this business any good? Are the problems temporary and not permanent? And does the management have a rational plan to restore the earnings?”

Temporary and not permanent: that is the stuck-versus-broken question I discussed earlier.

The research shifts the odds on the value bet:

“We try to find things where even if you’re 50-50 in your guess, you would win. And if you can be 60-40 because of an intense research process, then you can have a great record.”

Pzena does not claim any edge on the timing bet though. His average holding period is around three and a half years, and he is explicit that the waiting is structural:

“What makes something cheap is that you really don’t know what’s going to happen in the future... So we buy before we know.”

You cannot wait to see the ice forming and then buy the bottle; by the time the improvement is obvious, most of the stored energy has already been released.

What’s interesting looking at Pzena’s portfolios is that even with all that research, he still owns the population. Across all strategies, his firm holds 159 names, with the largest position at 6.5% of the combined portfolios. The US Small Cap Value Fund holds around 50 stocks, most sized at 2 to 3%. Even US Best Ideas, his most concentrated portfolio, holds 15 to 25 names, which is close to the 20 to 30 Greenblatt recommends for the Magic Formula.

Deep value in the Guru Gems portfolio

My own designated deep-value turnaround, Baxter (BAX), which I covered most recently in July, has been an interesting case study. On 30 July, Baxter reported organic revenue growth of 5% and adjusted EPS well above consensus, and raised both its top- and bottom-line outlook for the year. The stock rose over 15% that day, extending a climb of nearly 60% from its March low.

Baxter is the type of stock I have been describing here: statistically cheap, with problems that may or may not turn out to be fixable, and a timing question nobody can answer. It sits in my Guru Gems portfolio, and writing this issue did make me wonder whether it would be better placed in the Magic Formula basket. I have kept it where it is, because unlike a screen pick it is a name I researched in depth before buying, and one I have real conviction in.

Rich Pzena clearly has real conviction in it too. Baxter is the second-largest holding at 7.2% in his US Best Ideas strategy, his most concentrated portfolio. Even as one of his very best ideas, though, a stock like this gets a single-digit weight in a portfolio of 15 to 25 names.

Mine is at 5–6%, which feels about right for a stock where I am confident about the value and have no idea about the timing.

Closing: two lessons from metastability

Congrats, you made it to the end of this post :)

Or maybe you just scrolled to the closing part, in which case this one is for you. Two things to remember about supercooled stocks:

1. Being right about value is not the same as being right about timing.

A statistically cheap stock is a metastable state: real stored energy behind a real barrier. Your research can tell you whether the stock is stuck or broken, whether the normalized earnings still exist, whether the barrier is finite, etc.

It cannot tell you when the push arrives. You can be completely right about a stock and still look wrong for years while you wait for something to happen.

2. When timing is unknowable, own the population.

Boltzmann could never say which molecule would clear the barrier, but he could say precisely what fraction would, and that was enough to build a science on.

Greenblatt’s 20-30 stock Magic Formula basket and Pzena’s 50-name funds are the same move: convert an unanswerable question about one cheap stock into an answerable question about a group of cheap stocks.

Curate the population as carefully as possible. Just don’t bet the outcome on a single name!

Final thought

Both lessons above apply to stocks that look cheap.

This is not an argument against concentration in general. If you own a handful of businesses you understand deeply and that are compounding away, you are playing the game from Issue 2, where barriers and triggers are not really the point.

As you know, I run two portfolios: the Guru Gems portfolio holds a more concentrated set of names I have researched one by one, while the Magic Formula portfolio holds 25 ‘cheap-and-good’ ones on a one-year rotation.

But if you find yourself drawn to a stock mainly because it looks cheap, that is the moment to think about the distribution rather than the single name.

That’s it for this week!

I hope you enjoy these ‘Physics for Investors’ issues. Please give me feedback on how to make these even more interesting or what topics you would like to see covered.

If you enjoyed this post, I would really appreciate it if you could like or share it or leave a comment!

You can follow me on Substack @gurugems and X @guru_gems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

If you’re new to Guru Gems: I study the world’s best long-term investors and share what I learn, then apply it to two real-money portfolios. (Learn more here)

On the menu today:

• 📝 Seth Klarman update: his biggest Q2 buy

• 💎 Guru Gems Portfolio update: the winning and the losing side

• 👀 Watchlist update: three names I’m watching

Here is the 2-minute AI audio version:

🕵️ Seth Klarman’s special situation stock

Three weeks ago I wrote a post on Seth Klarman, the legendary value investor who runs Baupost and the author of ‘Margin of Safety’.

The post came just before the release of investors’ Q2 portfolios, so I highlighted a few names from Klarman’s Q1 portfolio, without going into too much detail.

Now that we have the latest 13F (US equity positions on 30 June 2026), I updated the post with the latest Q2 data, some additional commentary, and a closer look at Klarman’s biggest buy during Q2: Genuine Parts Company (GPC).

I analyzed why Klarman may have bought this company and whether I should consider adding it to my Guru Gems portfolio.

Here is the updated post:

Subscribe now

→ If you enjoy this newsletter, please subscribe and like this post 🤍

→ If you find value here, consider buying me a virtual coffee ☕

💎 Guru Gems portfolio update

The Guru Gems portfolio currently holds 14 positions and 22% cash. As of 30 August the portfolio is up 4.7% year to date and 9.3% over the past twelve months, against 12.4% and 19.4% for the S&P 500.

Three of my larger positions spent most of 2026 in a drawdown, and sitting on 22% cash in a rising market has not helped either.

For now, my Magic Formula portfolio (23% YTD, 25% 1Y) is the clear winner of my 2 portfolios.

Rather than walk through every holding, I want to focus on a few names where something changed this month.

Updates from the ‘winning’ side

Airbnb (ABNB)

Airbnb touched a new 52-week high of $193 earlier this week.

I started the position in June 2025 at $134, added in August ($123) and October ($125), for an average cost of $129. That is a gain of nearly 47% in a bit over a year.

The big move came after Q2 results on 6th August: revenue up 17% to $3.6 billion, nights and seats booked up 10%, adjusted EBITDA up 21%, and full-year guidance raised to at least mid-teens revenue growth. The shares jumped 17%.

Airbnb is currently trading at 43 times trailing earnings, which looks very expensive.

But as Bill Nygren argued when I covered him last year, that number overstates the case, because GAAP accounting pushes Airbnb’s long-term investment spending straight through the income statement. On a free cash flow basis it looks rather different. Trailing free cash flow of $4.8 billion puts the stock nearer to 23 times, which already looks more reasonable.

With the stock up almost 50% vs my average purchase price and the position getting close to 10% of the portfolio, I’m considering trimming if it keeps going up in the next weeks or months.

To close on Airbnb, I wanted to share a piece I came across which I really enjoyed.

(New Horizons Investment) writes a series of case studies where he analyzes companies through the Nick Sleep / Nomad Investment Partnership lens.

Airbnb was the first company in the series - check it out!

New Horizons Investment

Case Study #1 — Airbnb: A Marketplace Built on Trust

Airbnb Through the Nick Sleep Lens…

Read more

2 months ago · 5 likes · Chris Von Hoene

Guidewire Software (GWRE)

I added Guidewire as a 1-2% starter position in early July, after studying it following Bryan Lawrence’s portfolio review. It was one of the many software names dumped by investors on fears that AI would come and replace it all.

Lawrence had clearly done his homework and added for 3 consecutive quarters to his position.

The stock is now up 100% (!) from its June low and my position is up 44% since I bought it.

Unfortunately, at a 1-2% weight, a 40+% move doesn’t do much at the portfolio level.

This is the Greg Padilla lesson from Stock Master Maestros but in reverse: standard-sized starter positions protect you when you are wrong, but they also cap you when you are right.

Lawrence increased his own position by another 10% in Q2, on top of the 73% increase in Q1.

Updates from the ‘losing’ side

A few names have hurt this year. Most have started to recover a bit and my thesis for owning them hasn’t really changed.

Copart (CPRT)

Copart’s drawdown reached close to 60% at the July trough, the largest drawdown in more than 20 years, for a debt-free business with $4.2 billion of net cash.

The main cause is US insurance unit volumes falling 4.2% in the fiscal third quarter, as higher premiums pushed consumers toward liability-only policies and higher deductibles. Fewer claims → fewer totaled cars.

But I don’t believe the moat is broken:

• Total loss frequency reached 23.6%, nearly five points higher than four years ago. Repair costs keep climbing on sensor-laden cars, and strong auction returns make the total loss decision easier for carriers.

• Pricing power is intact. Even with volumes falling, average selling prices kept rising (mid-single digits globally, high-single digits for US insurance auctions). A broken business doesn’t get to raise prices into declining volumes.

• The balance sheet is a fortress. Roughly $4.2 billion in cash against essentially no debt, with ROIC around 29%. Morningstar still rates the moat Wide.

• Management is buying back stock. After years of hoarding cash, Copart repurchased about $1.6 billion of shares over the past year (~6% of market cap), with the stock below analysts’ fair value.

Jay Adair returned as CEO on 31 July, replacing Jeff Liaw. He used a special investor call on 6 July, the first between-earnings call in Copart’s 32 years as a public company, to describe the consumer retrenchment as cyclical rather than structural.

That is exactly what I wrote in my last update on Copart in June. The stock has recovered to around $33 from roughly $28 in mid-July, and fiscal Q4 results in September will probably be the real test.

On the Guru side, the conviction also remains with Peter Keefe increasing his position by 68% (!) during Q2, making Copart his 4th largest position at nearly 9% of the portfolio.

Finally, here is a recent piece from the team on Copart with lots of interesting charts:

Fiscal.ai

Copart: From Scrap to Scale

Hey there👋 Welcome to the first ever Fiscal.ai substack post…

Read more

a month ago · 15 likes · Fiscal.ai

PayPal (PYPL)

PayPal has been on a rollercoaster over the past months, and this week was no exception.

I bought in November 2025 at an average price of $59.90. By February the position was down nearly 40%, after a weak fourth quarter and a CEO change that took the shares down 20% in a day.

Then in July, Reuters reported that Stripe and Advent had offered $60.50 a share, valuing PayPal at $53 billion. The shares jumped 16% on the day and kept climbing, and the position went back into the green in early August.

On Friday, Bloomberg reported that the talks are off, with the two sides unable to agree on price. PayPal’s management considered $60.50 too low and the shares fell 11%.

I have no intention of selling here. A buyer with the money to do it valued PayPal at $60.50 a share and the board thought that was too low. Whatever else that tells you, it means two parties who had done the work both put the business above where it trades today.

Lindsell Train made a similar point in a recent letter: “Regardless, this initial price discovery exercise is a helpful reminder of how unduly cheap PayPal has become.”

In the meantime PayPal continues to repurchase its own shares, spending pretty much all of their free cash flow on it:

Constellation Software (CSU) & CCC Intelligent Solutions (CCC)

Two other names caught in the SaaSpocalypse selling are slowly recovering. Constellation Software is up 23% over the past six months and CCC Intelligent Solutions is up 34%. Neither is back to where it started, but the thesis that AI would quickly commoditize workflow software is looking less obvious than it did in February.

👀 Watchlist update

Following the latest 13F filings a few weeks ago, I updated the Guru Gems watchlist, which now includes more than 40 companies that are in Gurus’ portfolios and that may be interesting to study in more depth.

Here are three names I’m now watching closely.

1. Netflix (NFLX)

Netflix was one of the most-bought new names across the Gurus in Q2, with new positions for Bill Ackman, Terry Smith, Wallace Weitz and Tom Gayner (very small position), and Oakmark and Tom Russo adding on top.

Ackman’s move is the one that made the news, since he sold his entire Netflix position in a single day in April 2022 for a loss of around $400 million.

One name to add to that list is Marc Werres from Hinde Group (no 13F available), who started buying Netflix again in Q1 2026, which I covered when I wrote about him.

He built his first position into the 2022 collapse, held it through a 75% drawdown, and trimmed above $1,000 in 2025 for 4.1 times his money over 41 months. He is now rebuilding after the shares fell more than 30% from their June 2025 high on AI disruption fears and the failed bid for Warner Bros. Discovery.

The share price went as low as ~$70 during Q2 so that is the best possible price the Gurus could have bought it. It’s currently trading ay ~$82 and if it were to drop below $80 again I may consider starting a small position.

2. Moody’s (MCO)

A clear trend in Q2 was the accumulation of positions in so-called ‘toll booths’, businesses that charge a small, non-negotiable fee or transaction tax for every use.

Moody’s was one of them: Moody’s takes a small fee on every bond issuance and collects it with almost no incremental work.

What was particularly interesting was Bryan Lawrence opening a brand new position. He runs a concentrated portfolio and can go several quarters without buying anything at all, so a new position in his portfolio is always worth paying attention to (see also the Guidewire Software position discussed above)

The shares are already up 20% from their Q2 lows, but I think it’s still worth taking a closer look at this Guru favorite.

3. Watsco (WSO) — a new one

François Rochon opened a small new position in Watsco in Q2, the largest HVAC (Heating, Ventilation, and Air Conditioning) and refrigeration distributor in North America.

My first question was whether this was a data center play, since every HVAC name has been re-rated on the AI buildout this year. But it is not. Data center cooling is applied equipment sold largely direct by Trane, Johnson Controls, Carrier and Vertiv, while Watsco distributes equipment to contractors, and roughly 75-80% of it is residential.

Watsco operates as a middleman distributor selling to local, licensed contractors. They have 13% of a fragmented $75 billion market with more than 2,000 competing distributors left to acquire, a net cash balance sheet, and demand that is 70-80% replacement, which is about as close to non-discretionary as building products get.

That is the kind of business Rochon tends to own for a decade.

Why did Rochon buy it now?

Weak residential volumes and consumers choosing to repair rather than replace took Q2 revenue up only 2% to $2.1 billion and EPS down 12% to $4.00, and the shares are nearly 40% below their 2025 highs.

Morningstar’s fair value of $433 against a share price of $313 is essentially a bet that the replacement cycle returns, which is probably the same bet Rochon is making.

There are 2 key risks to watch. Founder and CEO Al Nahmad is 85 and succession is unresolved even with his son A.J. as president. And HVAC distribution is a local business with low barriers to entry, so the consolidation runway may be smaller than it looks.

I added Watsco to the watchlist and this could be an interesting ‘boring’ company to explore a bit further.

That’s it for this week…thank you for following along and I’ll be back next week.

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome to Round 11 of the Magic Formula Portfolio!

Last month marked the one-year anniversary of the portfolio and since then, every round is now both a sell and a buy.

Quick reminder if you are new to Guru Gems: while studying investing masters, I am creating two real-money portfolios to put my learnings into practice.

To learn more about the Guru Gems journey, start here.

Magic Formula Refresh

In ‘The Little Book that (Still) Beats the Market’, Joel Greenblatt suggests that a regular investor can beat the market averages by buying a group of ‘good’ companies at ‘bargain’ prices.

• ‘Good’ → High return on capital

• ‘Bargain’ → High earnings yield

Since he expects the magic formula to work on average, Greenblatt suggests to own a basket of 20-30 stocks, to buy this basket over the course of a year, and to hold each stock for exactly one year.

So that’s exactly what I’m doing…

You can read more on Joel Greenblatt and his Magic Formula here.

You can also download my Joel Greenblatt e-book for free:

Joel Greenblatt e-book by Guru Gems

494KB ∙ PDF file

Download

Download

One Year and One Month in: Current Magic Portfolio Performance

The Magic Formula Portfolio returned +26% since I started on 7 July 2025.

On a YTD basis, the Portfolio (+21%) is well ahead of the S&P 500 (+12%).

Two observations:

• The gap vs S&P500 has significantly widened since last month. In Round 10 the YTD numbers were +14% versus +9%. The portfolio has kept pace with no exposure to the AI trade that continues to drive the index.

• The winners are still not the obvious ones. I have said this before and it keeps being true. The biggest winners are names I would never have picked on conviction so I’m glad the formula picked them.

“If you are able to stick with the magic formula strategy through good periods and bad, you will handily beat the market averages over time” — Joel Greenblatt

Subscribe now

Selling Cohort 2

Greenblatt’s rule is to hold each stock for exactly one year, then sell.

This week I am selling the names I bought on 15 August 2025: Crocs (CROX) and Molina Healthcare (MOH).

The third name in this cohort, Catalyst Pharmaceuticals (CPRX), already sold itself. Angelini Pharma completed its acquisition of Catalyst on 16 July at $31.50 per share, against my purchase price of $20.00. That is a 58% return!

Cohort two ends with an average return of roughly 41%!

New Magic Formula Stocks

As I explain in detail in my first Magic Portfolio post, I use Joel Greenblatt’s Magic Formula screener combined with financial ratios data to identify the top stocks in terms of combined rank for Return on Capital (using ROCE) and Earnings Yield (using EV/EBIT).

I extend the analysis with additional criteria: P/FCF (an alternative valuation lens to EV/EBIT), Revenue 3yr CAGR (is the company still growing?) and Net Debt/EBITDA (how much debt is on the balance sheet?).

Altria (MO, Market cap $114B)

Magic Formula metrics: ROCE 61% (3yr: 60%); EV/EBIT 8.4x; Net Debt/EBITDA 1.4x

Altria is the dominant cigarette maker in the United States. Marlboro holds roughly 40% of the US market, against 14% for the next-largest brand. Beyond tobacco it holds an 8% stake in AB InBev and a 41% stake in cannabis producer Cronos.

It is also a nicotine pouch business. Through its wholly owned subsidiary Helix Innovations, Altria sells the On! brand into a category that has grown more than tenfold in five years to over $5 billion. Pouches now account for close to 60% of the oral tobacco market.

Why it is cheap

US cigarette volumes declined roughly 9% per year between 2020 and 2025. That is a real, secular decline, and probably why the stock trades at 8.4x EV/EBIT.

But volume is not revenue. Over the same five years, price increases blunted the revenue decline to about 2.5%. Over the past decade Morningstar estimates volumes fell around 7% annually while prices rose roughly 6% per year.

The pricing power has some room to run since cigarettes also remain cheap in the US relative to income. The World Health Organization (WHO)'s affordability measure (the share of GDP per capita needed to buy 100 packs of the most-sold brand) puts the US at about 1%, which is low against both developed and most developing nations. A pack costs roughly $10 in the US against $21 in the UK and $40 in Australia.

Morningstar rates Altria a wide moat, built on brand loyalty, regulatory entrenchment that keeps new entrants out, and the addictiveness of the product itself.

Build-A-Bear Workshop (BBW, Market cap $488M)

Magic Formula metrics: ROCE 29% (3yr: 33%); EV/EBIT 8.5x; Net Debt/EBITDA 1.2x

Swapping one of Michael Burry’s largest holdings (MOH) for one of his newest (BBW).

That is a coincidence rather than a strategy. Molina is leaving on the one-year rule, but I thought it was too neat not to mention it.

In his most recent update, Michael Burry described Build-A-Bear as “a remarkably simple and yet misunderstood business” and added it to his portfolio.

That is not why I am buying it though. The Magic Formula Portfolio is meant to be mechanical. BBW is here because it ranks well in the Magic Formula screener.

The business

I have to admit I had never heard of this company before, but I assume that if you live in the US, it is a well-known name. Most people know Build-A-Bear as the mall store where children stuff their own teddy bear.

Here are some key figures for this business:

• 376 corporately-managed stores plus 112 international franchised locations

• Gross margin of 57% and operating margin of 13%

• 5-year average ROCE of 32.5%, with 5-year ROIC at 24.2%

• Trades at 9.1x earnings

These are software-like gross margins, but priced as though it were an ordinary mall tenant.

The growth is coming from asset-light channels: third-party retail locations grew 22% over the trailing twelve months and international franchised stores ~17%, both of which expand the footprint without the capital cost of opening corporate stores.

Blue Bird Corporation (BLBD, Market cap $1.96B)

Magic Formula metrics: ROCE 32% (3yr: 47%); EV/EBIT 10.2x; Net cash

Blue Bird makes school buses. Headquartered in Macon, Georgia, it is one of the “Big Three” in North American student transportation, building Type C “Vision” and Type D “All American” buses in diesel, propane, gasoline and increasingly electric.

Some key figures:

• 3-year ROCE of 47%, 3-year ROIC of 44.5%

• Net cash (roughly $117M of cash against effectively zero net debt)

• Revenue compounding at 15.8% over five years, with operating income compounding at 16.3% against revenue at 7.6% over the longer run: margins expanding, not just volume growing

Why a bus manufacturer has a moat

Blue Bird’s advantages are more durable than I would have thought:

• Switching costs. School districts standardise on specific chassis and body configurations. Changing manufacturer means retraining mechanics and retooling parts inventory across an entire fleet.

• Distribution. An exclusive dealer network built over decades that a new entrant cannot replicate quickly.

• Aftermarket. A large installed base generating recurring, higher-margin parts and service revenue.

The electric transition is strengthening rather than threatening this. Blue Bird is a first mover in zero-emission school buses and has extended into Blue Bird Energy Services, selling charging infrastructure and energy management alongside the vehicles.

Key risks

• Policy dependence. A meaningful part of demand is tied to federal and state electrification funding, including the EPA’s Clean School Bus Program. A shift in appropriations would hurt.

• Customer concentration. Buyers are school districts funded by tax levies. Budget austerity means deferred fleet replacement.

• Returns are off their peak. Current ROCE of 32% sits below the 47% three-year average. The base period was depressed, so some of that average reflects recovery rather than a run rate.

Three names in, bringing the portfolio back to 25 stocks.

That’s it for this week! Please like or share if you enjoy reading Guru Gems.

You can follow me on Substack @gurugems and X @guru_gems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

It’s that time of the quarter when Superinvestors ($100M+) have to report their US equity positions held on 30th June 2026 (‘13F filings’).

I’m sure you have seen the headlines by now (“Ackman unveils six new investments including Netflix, Visa, Mastercard”; “Berkshire boosts Alphabet stake to third-largest holding”) and you have also been bombarded on X or Substack with screenshots from Dataroma showing individual portfolios.

What you haven’t seen however, is the analysis that distills the signal from the noise and identifies the hidden trends across the Gurus.

That is what I will try to share in today’s post.

• 👑 Alphabet: still the most-owned Guru Gem though also the most-trimmed

• 💾 Software buying continues, but it got selective

• 🛣️ The toll booths: Visa, Mastercard, Moody’s and S&P Global

• 👀 Multi-Guru accumulation: the names worth watching

• 🪄 What the Gurus think of my Magic Formula stocks

Let’s jump in!

New to Guru Gems? Start here

First, a detailed table listing most of the Gurus covered so far in the Guru Gems series as well as some other interesting Gurus.

The table shows the two largest positions in their portfolios (‘Gem 1’ and ‘Gem 2’) as well as the largest increase and largest decrease within their top 10 positions.

→ For the full table with more names and additional comments, access the Guru Master Google Doc, which also contains my full portfolio and latest activities

1️⃣ Are the Mag 7 still Guru favorites?

Alphabet remains the ultimate Guru Gem

Looking across the full table, Alphabet appears as Gem 1 or Gem 2 for more Gurus than any other stock. That has now been true for five quarters in a row.

• Li Lu holds it at a massive 48% of his US portfolio.

• It is also Francois Rochon’s largest position at 11%

• It is the second largest position for Bryan Lawrence (16%), Tom Gayner (9%), John Armitage (11%), Thomas Russo (12%), David Rolfe (9%)

• Berkshire significantly added, making it their 3rd largest position

At the same time, many Gurus (Oakmark, Russo, Rolfe, Armitage, Rochon) are trimming their Alphabet position, which is probably not surprising after Alphabet’s massive run over the past few years

Amazon is the clear second favorite

For the past quarters, Gurus were loading up on Amazon. AMZN was the laggard of the Magnificent 7, with barely a 1% return over full 2025 and a 7-8% decline over Q1 2026. This trend continued and more Gurus have been buying Amazon.

Seth Klarman made Amazon his #1 position in Q1 and continued buying in Q2 (+20%). Check out last week’s post where I unpack Seth Klarman’s investing philosophy and his thesis for owning Amazon:

David Rolfe added +35%, and ValueAct added a little bringing AMZN to 12% of the portfolio and making it their second largest position.

Not everyone was adding. Bill Ackman reduced by 25% (it remains a top-5 Pershing Square position), and Josh Tarasoff trimmed by 10%, though it remains his largest position at 18%.

Rest of the pack: META, MSFT, AAPL, NVDA, TSLA

Meta & Microsoft - Still Guru favorites

Apple - Many Gurus trimming

Nvidia & Tesla

NVDA and TSLA remain remarkably absent.

Except for John Armitage (NVDA 7%) and Josh Tarasoff (TSLA 7%) none of the other Gurus covered throughout the Guru Gems series so far have any positions in these 2 companies.

Subscribe now

2️⃣ Software buying continues, though more selective

In last quarter’s edition, I already highlighted how some Gurus were scooping up shares of companies hit by the SaaSpocalypse.

Q2 shows continued software buying, but more selective.

Bill Nygren at Oakmark continued to add to Salesforce (CRM), making it now a top 3 position. They also added to Adobe (ADBE, 1%) just like Samantha McLemore at Patient Capital Management.

Akre Capital continued adding to ServiceNow (NOW, 2%), a position they initiated in Q1. They also own Salesforce (CRM, 2%) and CCC Intelligent Solutions (CCC, 3%).

Finally, Bryan Lawrence (Oakcliff Capital), who rarely makes changes to his highly concentrated portfolio, increased his Guidewire Software (GWRE) position by 10%, following his 73% increase in Q1.

Lawrence views Guidewire as an ‘AI-Resilient’ software company, and this addition seems to double down on that conviction. I decided to add GWRE to the portfolio last month, after studying it in more detail:

3️⃣ Multi-Guru accumulation: the names worth watching

One of the most useful signals in a 13F analysis is when the same name appears in multiple portfolios at the same time. This quarter there are a few that have seen some significant Guru buying.

🛣️ The toll booths: Visa, Mastercard, S&P Global, and Moody’s

One clear trend in Q2 was the accumulation of positions in so-called ‘toll booths’, businesses that charge a small, non-negotiable fee or transaction tax for every use.

Visa (V) & Mastercard (MA)

• Bill Ackman initiated both (5-6% each), two of his six new positions

• Visa is Gem 1 for John Armitage (14%, +4%) and ValueAct (14%), and Gem 2 for Chris Hohn (20%), making it the second most-owned name in the whole table

• Mastercard is Akre Capital‘s largest position at 20%, though they trimmed it by 13%

• Terry Smith did both at once: Mastercard was his second-largest new buy, funded in part by cutting Visa, a position he has held for many years

• Dev Kantesaria cut MA by 28% (his largest decrease) but it remains his 4th largest position (18%).

S&P Global (SPGI) & Moody’s (MCO)

• SPGI was a new position for Ackman (5%), Triple Frond Partners (4%), Chris Bloomstran (2%) and Armitage (1%), plus a significant increase for Pat Dorsey (8%)

• Chris Hohn holds both MCO (12%) and SPGI (11%), both unchanged

• Dev Kantesaria also holds both: SPGI (20%, trimmed -13%) and MCO (19%, unchanged)

• MCO was a brand new position for Bryan Lawrence

Bryan Lawrence’s move is worth watching a bit more closely. When a concentrated investor who does nothing for quarters at a time opens a brand new position, it’s worth paying attention.

Netflix (NFLX)

Netflix was an equally popular name in Q2.

• New position for Bill Ackman (5%), Terry Smith (4%), Wallace Weitz (1%), and Tom Gayner (0.2%)

• Both Oakmark (1%) and Tom Russo (5%) added to their position and Josh Tarasoff still holds it 5% after he bought it in Q1.

Bill Ackman getting back into Netflix was big news.

Earlier this year I wrote about the lessons learned from Ackman’s first Netflix episode:

Marc Werres, a lesser-known Guru, also bought NFLX in 2022, at roughly the same time Ackman did. Werres made 4x his money after selling most at the highs in 2025.

In Q1 2026 Werres started buying again.

Read the full story in my post on Marc Werres:

Here are two excellent deep-dives I’m currently going through to gain a better understanding of the business: ’s Netflix ($NFLX) Deep Dive and ‘s Netflix ($NFLX): Down 50%. Is it Finally Cheap?

Uber Technologies (UBER)

UBER is also worth highlighting:

• It is Bill Ackman’s largest position and he added more during Q2

• Pat Dorsey significantly increased his position - now 8% of his portfolio

• Finally, UBER was one of the 12 new names Terry Smith added to the Fundsmith portfolio as part of his ‘Extreme Makeover’ (check his semi-annual letter if you haven’t yet). It’s the largest of the new positions and is now a top-5 of his US equities.

🪄 What about the Magic Formula stocks?

I thought it might be interesting to also have a look at the stocks in my Magic Formula portfolio (a portfolio of 25 names built on the principles laid out in Joel Greenblatt’s ‘The Little Book That Beats the Market’)

Small Caps

There are currently 8 Small caps in the Magic Formula portfolio, and only 1 of them appears in the portfolio of a Guru: Perdoceo Education (PRDO) is a very small position (0.3%) for Miller Value Partners.

This is not surprising as many of these companies are too small to own for larger funds.

In his book, Joel Greenblatt had already observed that the Magic Formula worked best for smaller-and-medium-cap stocks:

“It seems clear that there is a greater opportunity to find bargains in the small-cap arena both because there are more stocks to choose from and because smaller stocks are more likely to be lightly analyzed and, as a result, more likely to be mispriced.”

Mid Caps

There are also 8 Mid Caps in the portfolio and same story here, only 1 name comes up: Crocs (CROX) - the ugly shoe company.

• Norbert Lou holds it at a massive 27% of his portfolio

• Samantha McLemore, Li Lu, and David Einhorn all have smaller positions

CROX is one of my best performing stocks in the Magic Formula portfolio, currently up 57% in less than a year.

Large Caps

Finally, 9 large caps, and probably not surprising that 8 names out of the 9 appear in Guru portfolios: MOH / DECK / LULU / CTSH / CI / ADBE / BKNG

2 of them are worth highlighting:

Cognizant (CTSH)

Cognizant is one of the world’s largest IT services firms, and was cheap for the same reason many other service firms have been hit: the market fears AI will commoditize what IT-services firms sell.

I added Cognizant in the latest round in early July, and it is already up 40% (!) since I bought it, partly driven by a solid earnings report a few weeks ago.

• Rich Pzena has owned CTSH since 2024 and significantly increased his position over the past 2 quarters

• Chris Bloomstran initiated a new position during Q2

Booking Holdings (BKNG)

I added Booking Holdings, the world’s largest online travel agency, in early June.

The stock was down more than 25% over the past year, giving it a valuation that put it at the top of the Magic Formula screener: 3Yr ROCE: 72%; EV/EBIT: 13.6x

In my update I already highlighted that this was a ‘Magic Formula meets Guru Gems’ stock with François Rochon, AKO Capital and David Rolfe all owning it.

Interestingly AKO Capital sold their full position in Q2, but Bill Nygren started a very small position (0.4%) for Oakmark.

The stock is up 27% in the Magic Formula portfolio.

That's it for this week's edition!

A lot to digest from a single round of 13F filings, but a few clear trends came out of it, and a handful of names worth looking into.

For me, the two I’ll be watching most closely are Netflix and Moody’s. Both had multiple Gurus buying in Q2, and both could end up in the Guru Gems portfolio if I can get comfortable with the business and the price. I’ll keep you posted.

You can follow me on X @guru_gems and Substack @gurugems for more insights.

Until next week!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

In today’s edition:

👤 Guru in the spotlight: Seth Klarman
🛡 The six things Baupost does to protect capital
🫧 What Klarman says about today's AI market
💎 Gems in focus: Klarman's three largest positions

Today’s post is a bit longer, so grab a coffee and make yourself comfortable! If you’re in a rush, here is a 2-minute AI audio version of today’s post:

New to Guru Gems? Start here

👤 Guru in the spotlight: Seth Klarman

Seth Klarman has been running the Baupost Group in Boston since 1982. In those 44 years he has had five down years, and the worst of them was ‑12%.

This is probably not a coincidence when your whole investment philosophy is built around ‘downside protection’.

Regular readers will remember I shared some thoughts from Klarman’s June interviews in issue #63, and promised a full deep-dive.

Baupost’s origin story is an interesting one. In 1982 four Boston families were sitting on the proceeds of some business sales: a stake in Channel 5, the ABC affiliate, and a computer consulting and publishing business among them. Jordan Baruch, Isaac Auerbach, William Poorvu, and Howard Stevenson pooled $27 million and formed an expanded family office to protect and compound that wealth across generations. The firm’s name is an amalgam of their surnames (BA-AU-PO-ST).

They needed someone to do the actual investing. Poorvu had taught real estate at Harvard Business School, and he recruited a former student: a 25-year-old whose entire track record was a stint as an analyst at Max Heine and Michael Price’s Mutual Shares, the legendary value shop, between Cornell and his MBA.

So even though Klarman was not the founder, he was there from day 1, and hired to run it. The families were very clear about what they expected from him:

“Our mandate wasn’t really to grow an investment business. It was to deliver good returns with limited downside to those clients.”

Baupost today runs roughly $22 billion across public equities, credit, private investments and commercial real estate, and it still behaves like a family office that would prefer to miss an opportunity than lose the money.

Klarman wrote Margin of Safety in 1991, when he was 34. It sold badly, went out of print, and then became one of the most sought-after investment books ever written.

Over the past four decades, Klarman gave few interviews or talks in public. His explanation was that if you are accumulating a stock, you get a better price when nobody knows.

Over the past few years, it seems Klarman has been doing more interviews and I have gone through all of them to build this post.

Subscribe now

If you enjoy this newsletter, please help others discover it by clicking 🤍

And if you would like to support my work, consider buying me a virtual coffee ☕

🛡 The Baupost Way

When Klarman was asked in a recent interview what makes Baupost unique, his answer was all about how the firm avoids losing money:

“The way we protect capital is we focus on downside as much as upside. We do meticulous fundamental research on every company. We will hold cash in the absence of immediate opportunity. We don’t leverage the portfolio at all. We’ll buy senior securities… so that we’re structurally senior in a lot of what we do, which protects on the downside. And finally, we do macro hedges.”

Klarman mentions six distinct practices to protect capital in a single breath, so let’s unpack this a bit more and go through them one by one.

1. On downside

You can see this in the way Baupost approaches new ideas for the portfolio. When an analyst pitches him, Klarman wants to run the post-mortem before the investment has been made:

“If we’re wrong and three years from now we’re sitting here… going over what went wrong, what are we going to be talking about?”

Readers of my Chris Hohn deep-dive will recognise this mindset. Hohn also applies a risk-first framing to every investment: “The vast majority of investors focus on return. But in my career I focused on risk.”

The key lesson from these two brilliant investors is that you should decide what you can lose before you get excited about what you can make.

2. On doing the work

Klarman describes Baupost’s research process as “a mile wide and then a mile deep”: a team of generalists scanning everything, then drilling into a single name until they know it unbelievably well. The point of all that depth is what the work buys you on the day the market turns against you:

“The challenge is the discipline to wait patiently until those bargains emerge, and then the temerity to stand there when it’s going down in your face. The market’s telling you you’re wrong, but you’re holding on because you’ve done the work.”

Months of that kind of work create their own problem, though. The analyst who went a mile deep does not arrive at the meeting neutral:

“The analyst needs to fall in love with the positions. As portfolio manager, I just date their ideas.”

To counter this bias, Klarman keeps asking his analysts questions. If we didn’t own this already, would we still buy it today? And before buying a single share, he wants to know where they would sell:

“If they say, ‘Well, I like it at 70, but I’d sell it if it got to 80’… it’s like, you’re that sure that it’s going to move 15 percent? Because what if it goes to 50 before it goes to 80? Maybe it’s not the right entry point.”

3. On cash

Baupost holds cash, sometimes a great deal of it, and Klarman has spent decades defending the practice against clients who reasonably ask what they are paying him for. Here is what he said on this during a lecture to a Columbia MBA class in 2010:

“To say you have to look only at today’s opportunity set, and therefore be fully invested all the time, is sort of like saying you have to marry someone from your high school graduating class as that’s the whole opportunity set.”

Although this is also the one guardrail he has since walked back. Baupost ran 30% cash and higher at times, and he told in an interview that he “almost certainly made a mistake”, because the optionality never paid for itself across a decade of suppressed rates and no serious drawdown.

So he changed the mechanism rather than the principle. In his words:

"We made our liquid books more liquid, especially our public equity book, where we used to own companies with 500 million or billion dollar market cap. Now we have much bigger market cap holdings on average… we can pivot on a dime with a large percentage of our book. So we don't need as much cash."

More on his portfolio later in this post.

4. On leverage

Klarman is very clear on what he thinks about leverage:

“If we can’t make money unleveraged, we’d rather go out of business.”

And one of my favourite lines, which should be an important reminder for anyone using leverage:

“The risk of leverage is that it’s like drowning in a pond that, on average, is one foot deep.”

Leverage is what turns investors into forced sellers, and Klarman is explicit that "you never want to be a forced seller, and you especially want to be able to buy from forced sellers in any asset class."

5. On seniority

A large part of Baupost’s book sits above the equity in the capital structure: senior debt, structured private deals, instruments with contractual claims. Credit is, according to Klarman, the firm’s best discipline.

“Bonds are a wonderful safe instrument in the following context: you get senior securities, you’re first in line, contractual terms, you get interest, you get your principal back, you know when you’re getting those things back… The problem is everyone knows that, so they don’t return very much, most of the time.”

Most of the time, safety is priced in and returns are dull. But in distress, when downgrades and defaults force other holders to dump, Klarman gets to buy equity-like returns wrapped in bond-like protections.

6. On hedging

Finally, an overlay of macro protection bought at the least intuitive moment:

“We’re buying macro protection when vols are low and people think nothing bad is going to happen. So we can sell that at a gain… during a storm.”

The logic is the same as with any insurance: it is cheapest when nobody thinks they will need it. And these hedges do more than soften the blow. When the storm arrives, Baupost can sell them at a profit, which frees up cash at exactly the moment the best bargains start to appear.

All six guardrails end up serving the same purpose, which is to make sure Baupost is still buying on the worst day of the cycle.

The guardrails at work: 2008

When Bear Stearns and then Lehman came apart in 2008, Baupost, closed to new money for much of its history, called up its waiting list and raised several billion dollars inside a single quarter. The team deployed it at a reported pace of roughly $100 million a day, buying residential mortgage securities, the debt of the auto finance arms, and "very interesting" pieces in the collapsed Lehman capital structure.

While much of the industry sat paralysed, Klarman described his team as “kids in a candy store.”

But when Klarman was asked about this period, he pushed back on the heroic version, and clarified that it was all very calculated:

“The way you’re conveying it probably comes across as we come in with giant satchels of money and hand over fist deploy it. It wasn’t like that at all. It was the same cerebral, methodical, painstaking environment that we do every day… I can buy this bond at 70 that I think is covered at par.”

🫧 What Klarman says about today’s AI market

When asked about today’s market in a podcast with Morgan Stanley’s Michael Mauboussin, Klarman called the current environment (Aug 2025) “a 4 out of 10 opportunity set.” Markets at all-time highs, very little despair, very little panic.

When asked in another interview more recently (Jun 2026) whether this is a bubble, he would not go all the way:

“I think it has characteristics of a bubble… you certainly see it in places like when Allbirds, the shoe company, added AI to their name and the stock did well. And that’s crazy. That’s reminiscent of the dotcom.”

(Note how the thumbnail below says “THIS IS A BUBBLE”….not really what he said)

His real concern is about which way the multiple is moving:

“If you threw this much uncertainty into a market, I think you’d say maybe the market should actually be at a lower multiple to accommodate all this uncertainty. And instead the multiple keeps going up.”

But a low score is not zero, so it means there are still opportunities. Klarman splits the market into three buckets:

There are AI winners, which everybody wants. There are AI losers, which everybody is dumping. And then there is a third group he calls AI agnostic: businesses the technology simply will not affect much, which nobody finds exciting.

Baupost is spending its time in that third bucket, and picking through the perceived losers for anything that has been miscategorised.

As for the AI model companies themselves, Klarman is uninvolved. He argues that they have to spend enormous sums continuously just to stay in front, and if they ever fall behind they are in real trouble. For Klarman, that is not the description of a great business.

Which brings us, finally, to what he actually owns.

💎 Baupost portfolio

Below is an overview of Baupost’s top 10 US equity holdings at the end of Q2 2026.

This $5.4 billion, 23-stock portfolio is not all of Baupost. It reflects the 13F, which captures US-listed equities only, roughly a quarter of the firm. As highlighted in the introduction, Baupost’s total assets is roughly $22 billion across public equities, credit, private investments and commercial real estate.

Here is a closer look at the top 3 positions from end of Q1 [Note: my original post was written before Q2 holdings were released. I have now added some Q2 commentary]

1. Amazon (AMZN) — Q1: 12.7%; Q2: 16.5%

Amazon is Klarman's largest disclosed equity position, and he built it over just 2 quarters.

He opened it in Q4 2025 at 9.3% of the portfolio, then added another 47% in Q1. Going from zero to a 12.7% position in two quarters is quite a statement of conviction.

Amazon was trading in the $200-250 range during those quarters, so with a current share price of $274, he is probably up somewhere between 15%-30% on this position.

A position this size in a mega cap is also exactly what Klarman described when he talked about the cash decision. This is one of the “much bigger market cap holdings” (together with Alphabet) that lets Baupost “pivot on a dime with a large percentage of our book”.

As for the business itself, what Klarman likes is that Amazon is an enormous cash flow machine, versatile enough to play several roles in AI, with a data centre business growing much faster than anyone expected a year ago.

Klarman saw an opportunity to buy this wonderful businesses at an attractive valuation:

“You don’t have to buy them this second… there are blips in almost every company where you get windows of opportunity.”

It’s always pleasing to see when a Guru confirms one of my earlier picks. I bought Amazon in February at $204.50 and added in March at $207.40, for an average of $205.20. My Amazon position is up 33%, putting it in the top 5 best Guru Gems performers. (Alphabet, the ultimate Guru Gem, remains by far my best performer)

Klarman and the Guru Gems portfolio are in good company: Josh Tarasoff, Bill Ackman, Dennis Hong, and John Armitage, all have significant positions in Amazon, with the last three all adding during Q1.

2. Restaurant Brands International (QSR) — Q1: 11.7%; Q2: 9.0%

QSR is the franchisor behind Burger King, Tim Hortons, Popeyes and Firehouse Subs; it’s a capital-light royalty stream on other people’s restaurants.

Compared to Amazon, Klarman’s second-largest position is the opposite story: he has been slowly adding to this position almost every quarter since end of 2024.

The stock has been trading in the $65-$75 range for most of that time so depending on his entry prices, this position is probably a modest gain until now, with an additional ~3.5% dividend yield on top.

Just like Amazon, QSR is also a large position in Bill Ackman’s portfolio. Below is an extract from Ackman’s February 2026 presentation detailing why he holds QSR.

If you want to learn more about Bill Ackman and his Pershing Square approach, check out my post from February:

Long-time readers know I also have my own history with this stock.

QSR has been on the Guru Gems watchlist since my Chris Davis deep-dive back in issue #7, and it earned a Silver medal on my Podium of Errors in March as an error of omission. The stock came very close to $60 in September, I waited for it to go a bit lower, and it never did.

With Davis, Klarman and Ackman all holding meaningful positions, this remains a highly Guru-endorsed name on my watchlist, and if it ever revisits those levels I won’t make the same mistake twice.

3. WESCO International (WCC) — Q1: 7.7%; Q2: 4.2%

A company I had never looked at before, and the hardest of the three to write about.

WESCO distributes electrical, networking, security and utility equipment, the unglamorous plumbing of construction sites, industrial plants, power lines and, increasingly, data centres.

It doubled in size by acquiring Anixter in 2020, and the data centre business (part of Communications & Security Solutions in blue below) has been growing at roughly 50% a year to more than 20% of revenue. Second quarter organic sales grew 13% and the backlog grew 60%.

Klarman opened the position in Q1 2024 and then built it aggressively: he almost quadrupled it in Q2, added another 57% in Q3, and kept buying through 2024 and into the first half of 2025.

We don’t know his exact entry prices, but his average cost is probably somewhere around $165.

The stock closed last week at $364, so that’s well over a double.

Klarman has already started selling. In Q3 2025, with the stock around $211, he cut the position by almost a third. He trimmed a bit more in Q4 at $245. This is exactly what he says he does:

“[We protect] on the downside as best we can, by doing deep fundamental analysis, by knowing our names unbelievably well, by not being afraid to sell them when the price is up. And the same way, same as we buy more when the price is down.”

Morningstar raised its fair value estimate for Wesco in July, but only to $275, roughly 25% below where the stock trades today. It rates the moat narrow and says that "the market seems to be extrapolating Wesco's supercharged organic sales growth too far into the future."

It will be interesting to see from the Q2 filings (reflecting holdings on 30 June) how much Klarman will have trimmed this position. Or does he see something in the data centre buildout that others are not seeing yet…

Q2 Update: I ended the section on WESCO wondering how much Klarman would trim. The answer is more than half: he cut the position by 54%, from 7.7% down to 4.2%. If we add up the Q3 2025 cut, the Q4 trim and this one, he has now sold roughly 70% of the position. It seems Klarman might agree with Morningstar's $275 fair value estimate.

💎 The Gem: Genuine Parts Company (GPC) [Q2 update]

Klarman’s biggest buy of Q2 was Genuine Parts Company (GPC), the company behind NAPA Auto Parts. He added 1.3 million shares, an increase of 89%.

Background

On 17 February, GPC announced it was splitting itself into two separate companies. On the same day it also reported a weak quarter (adjusted EPS of $1.55 against $1.79 expected, North America Automotive EBITDA down 14%), and guided 2026 lower.

The market did not take it well and the stock fell 19% in a week. It kept sliding through the next months, and bottomed around $91. In February it was trading at $146.

“[We protect] on the downside as best we can… by not being afraid to sell them when the price is up. And the same way, same as we buy more when the price is down.”

I used the above quote for the selling in WESCO, and Genuine Parts is clearly the second half of the quote.

Genuine Parts Company — To buy or not to buy

As always, I’ll evaluate this Guru Gems candidate using my Terry Smith-inspired framework:

1/ Buy good companies

GPC consists of two separate businesses:

• Global Automotive (‘NAPA’, ~60-65% of revenues): distribution of automotive replacement parts through 10,000+ locations in North America, Europe and Australasia, two thirds of them independently owned. Around 80% of sales go to professional mechanics who need the right part ASAP, which is where the distribution network earns its keep.

• Global Industrial (‘Motion’, ~35-40% of revenues): second-largest industrial distributor in North America, with leading share in bearings and power transmission. Roughly $9 billion of revenue and more than 200,000 maintenance and repair customers.

Why GPC made it into the Baupost portfolio

Let’s look at GPC through Klarman’s eyes and see why this became his sixth-largest position.

• A special situation. The separation is a ‘tax-free spin-off’ targeted for Q1 2027.

  The company’s aim is to “unlock shareholder value by allowing each entity to pursue tailored strategies, optimize capital allocation, and enhance investor clarity”.

• The market sold the stock for a different reason. The spin and the bad quarter were announced on the same morning. Investors reacted to the earnings miss and the lower guidance, and the separation was buried underneath it.

• At $91, the downside was already in the price. At that price the stock was on around 8x EV/EBITDA, under 12x this year's adjusted earnings, and a dividend yielding 4.7% backed by 70 consecutive years of increases. That was the equity version of Klarman’s line: "I can buy this bond at 70 that I think is covered at par."

• Industrial was practically thrown in for free. We only learned this in July, but if the automotive half is worth what O’Reilly later offered, then at $91 Motion was valued at roughly 7x pro forma EBITDA. Applied Industrial Technologies trades near 20x.

• Nobody wanted it, which is the point. Remember Klarman’s third bucket: companies AI will not affect much, and which nobody finds exciting. Selling bearings to factories and brake pads to mechanics definitely fits that category.

So Klarman wasn’t probably paying up much for the separation. He was buying at a price the market had set on a bad quarter, and getting the catalyst for almost nothing.

Then, two days after the quarter closed, he got his confirmation. On 2 July Bloomberg reported a roughly $10 billion cash bid from O’Reilly for the automotive business, and the stock jumped 13% in a day. Klarman had already built the position by then. Three weeks later, CEO Will Stengel said GPC is “not currently in discussions with any competitor.”

This is not a high-quality compounder

I think it’s important to highlight that Klarman isn't buying a compounder here, he's buying a special situation, and that's how it should be judged.

Group EBITDA margin is 8.3%. Return on invested capital is somewhere around 10–12%. Organic growth is around 1–3%, with most of the top line bought through acquisitions (over 50 in 2025 alone).

Finally, Morningstar rates the moat ‘narrow’.

2/ Don’t overpay

At around $135, GPC trades at:

• ~17x this year’s adjusted EPS guidance of $7.50–8.00

• ~11x EV/EBITDA

• a 3.2% dividend yield

That’s a fair price, not particularly cheap. Since the separation is the whole thesis, it’s also worth doing the sum-of-the-parts.

Management shared the following in July: after allocating corporate costs and dis-synergies, each business lands at roughly $1.0 billion of pro forma EBITDA, which adds up to about the same $2.0 billion GPC earns today. The split creates no new EBITDA, so whatever value it unlocks has to come from the multiple.

Assume we hold the automotive half at O’Reilly’s $10 billion and the market is now paying about $13 billion for Motion, or roughly 13x. Then Motion has gone from about 7x at the lows to about 13x today. If we put it at 16x the stock is worth about $157; and at 11x, about $120. At $134 we are sitting in the middle.

3/ Final decision

The usual question with a Guru buy is what they see that the market doesn’t. Here I think we already know, and it isn’t really about the business. Klarman was handed a special situation at a price set on a bad quarter, and he took it.

Which leaves the question: what is left for the rest of us? Not much. Motion has re-rated from roughly 7x to 13x, the O’Reilly bid is public, and the price now assumes a deal management says it isn’t negotiating. The catalyst is still coming, but I would be paying full price for it rather than getting it for free.

📌 I’m passing on Genuine Parts. It will be interesting to watch what happens during and after the investor days, and what further moves Klarman will be making.

That’s it for this edition!

Let's close with another great quote from this brilliant Guru. I've really enjoyed studying Seth Klarman.

“People think they’re hiring a manager to make them money. But probably, they’re hiring a manager to keep them out of trouble and maybe fight their own instincts sometimes.”

As always, thank you for reading and following along on my journey. You can find me on X @guru_gems and Substack @gurugems for more Guru Gems insights.

Until next week!

Disclaimer: This newsletter is for informational purposes and does not constitute financial or investment advice. Always do your own research before making any investment decision.

This week's edition is a short one as I am taking some time off. But there is still something to celebrate, plenty to read, and a good investor letter to unpack.

• 🎉 Guru Gems is closing in on 500 subscribers: I’m giving away prizes!

• 📚 Summer reading: the most popular posts from the Guru Gems archive

• 📬 A quick review of Giverny Capital Asset Management's Q2 letter

🎉 Almost 500 of You

When I published the first Guru Gems issue back in April 2025, I would have been delighted with 100 readers. We’re now closing in on 500 subscribers (1150 followers on Substack!), and I’d like to celebrate that properly.

How to win (between now and subscriber #503, or 15 August at the latest):

🏆 Existing subscribers: the 2 readers with the most referrals win. Get your personal referral link here: gurugems.org/leaderboard

Share

🎲 New subscribers: everyone who subscribes before we hit #503 is entered into a random draw for the third prize.

Winners choose which of the three publications they’d like.

New here? Guru Gems studies the world’s best investors and applies their lessons to two real-money portfolios.

Why 503? It’s a nice prime number. The physics nerd in me occasionally leaks out :)

📚 Summer Reading

August is usually a good time of the year to catch up on some reading. Here are the posts I’d point you to, especially if you’ve joined recently and are wondering where to start.

My 3 most popular Guru deep-dives

• Norbert Lou — The Punch Card Mindset: why less is more, and how focus drives extraordinary returns. Also where the idea for the PayPal position started.

• Dev Kantesaria & The Valley Forge Way: extreme concentration, a medical doctor’s temperament, and FICO as the ultimate toll booth.

• Sir Chris Hohn & The TCI Way: fortress businesses, and why S&P Global sits in two of the world’s most disciplined portfolios.

My 3 favorite Gurus

• Bryan Lawrence & The Lawrence Method and, one year later, The Lawrence Method Revisited: six questions, being wrong 30% of the time, and a beaten-down software name.

• Timeless Lessons from François Rochon: the Rule of 3, the Podium of Errors and The Giverny Way: 32 years of compounding with patience, humility and rationality

• A Hidden Master: Marc Werres & The Hinde Way: seven positions, two years without a new idea, and one of the best letters almost nobody reads.

The quarterly Guru portfolio updates

Every quarter I go through the 13F filings of the most interesting Gurus and try to separate signal from noise. Here’s the most recent one: What Gurus are buying and selling — Q1 2026.

The Q2 update will land in two weeks.

And if you still want more to read…

…check out the full Guru Gems Archive:

Subscribe now

→ If you enjoy this newsletter, please subscribe and like this post 🤍

→ If you find value here, consider buying me a virtual coffee ☕

📬 A Good Letter: Giverny Capital Asset Management, Q2 2026

Regular readers know I always enjoy a good investor letter. I wrote about my three favourites a while ago:

David Poppe’s quarterly letter is certainly also one I always look forward to.

Quick context: Giverny Capital Asset Management (GCAM) is David Poppe’s New York firm, run in partnership with François Rochon’s Giverny Capital. Poppe previously ran Ruane, Cunniff & Goldfarb, the firm behind the Sequoia Fund.

So his letters are written from deep inside the quality-compounding tradition.

GCAM 2026 Q2 Letter

405KB ∙ PDF file

Download

Download

Three takeaways from the Q2 letter:

1. 🗳 The market is voting, not weighing

Poppe reminds us of Graham’s famous line: in the short run the market is a voting machine, in the long run a weighing machine. Right now, he argues, the voter has some fairly extreme views.

The S&P 500 returned about 10.2% in the first half, and yet 210 stocks in the index were down, while roughly 200 beat the index return. A 40/60 split isn’t unusual. What is unusual is that two-thirds of the laggards trailed by more than 10 percentage points.

That dispersion would make sense if the earnings were concentrated in the winners, except they aren’t.

We own Index constituents such as Charles Schwab, JP Morgan, Mastercard and Progressive Corp. that continue to grow their earnings per share at healthy rates and in some cases at higher rates than usual, but whose stock prices are lagging.

Poppe also highlights Constellation Software, which kept compounding adjusted EPS and free cash flow per share at roughly 20%, while the stock dropped 21%.

The market speculates that CSU’s earnings may grow in the near term, but over time Artificial Intelligence (AI) applications are going to enable lower-cost competitors to enter the business.

All of this is to say, we’re out of sync with the Index and we’re not inclined to change course. I believe we own high quality earnings compounders that will benefit from implementing AI capabilities and are likely to maintain their dominant competitive positions. We’re generating a good return, but I think our companies overall are performing better than their stock prices would indicate.

2. ⚡ Momentum has taken over

Poppe highlights a very interesting chart in his letter which counts the number of days per year the US Momentum Index moves up or down by at least 1.5%.

A decade ago: roughly 10 days a year. In 2020: 71. Annualised for 2026: 70, the most volatile since COVID.

Poppe writes that the market over the past few years increasingly has become skewed to chasing momentum and believes this is driven by “a push and pull between fear of missing out on a revolutionary technological advance and of being wiped out if the advance ends up delivering disappointing financial returns.”

He illustrates what that feels like from inside his portfolio: on a single day in May, one GCAM holding (AAON) reported strong earnings and rose 31%, while another (Installed Building Products) reported mildly disappointing earnings and fell 28%.

3. 🚀 Betting on moonshots

On the bubble question, Poppe sees some red flags, but also acknowledges that “Corporate America is remarkably efficient, with profit margins and returns on capital far higher than they were a generation ago”

The market is strong for good reasons, yet this feels like a moment where people are betting on moonshots.

The moonshots may well happen. But what if they don’t happen soon? In the meantime, some very good companies are being left on the floor.

He flags a risk we’ve also heard from Paul Tudor Jones, Seth Klarman, and David Rolfe at Wedgewood: the equity supply/demand balance is about to change drastically:

Massive IPOs either completed or on tap from AI darlings like Space Exploration Technologies, Open AI and Anthropic, plus chunky secondary stock and/or debt offerings from Alphabet, Amazon, Nvidia, Oracle and others may eventually drain what has been a very deep pool of investor liquidity.

📊 GCAM Portfolio Activity

Alphabet (the ultimate Guru Gem) and Arista Networks are GCAM’s largest positions, even after trimming both 20% and 17% respectively. Both positions were in the top 5 contributors to performance over the first half of the year:

My two cents on the best and worst contributors is that our top five are terrific, have run up a lot this year and now are expensive. Our bottom five are terrific, unpopular with Mr. Market and, as a result, are attractively priced.

They used the proceeds of the Arista and Alphabet sales to add to their financial stocks:

Charles Schwab ($92 and $85), Kinsale Capital (mid-$330s), TWFG (~$19) and Progressive ($205)

🔮 Next week

I’m doing a lot of reading during my break, partly for pleasure, and partly to finish a Guru deep-dive I’ve been working on for a while: Seth Klarman.

I finally managed to get my hands on a copy of Klarman’s iconic ‘Margin of Safety’, so this will hopefully be a nice add to the research I’ve already conducted.

Stay tuned!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

In today’s update:

• 💎 Guru Gems Portfolio — a mid-year look in the mirror

• 🔍 Alphabet’s phenomenal, historic, confusing quarter

• 📝 A quick word on⚕️ Baxter and 💸 PayPal

• 📜 Wedgewood Partners: Are Hyperscalers still Magnificent?

Start here if you are new to Guru Gems.

2-minute AI summary of today’s post:

💎 The Guru Gems Portfolio at Mid-Year

As I have written before, I’m all about transparency with Guru Gems, even when the numbers aren’t pretty.

The Guru Gems portfolio is up just 2.2% since inception (Apr 2025). Not exactly a wonderful result, and it stings even more considering my biggest position, Alphabet, is up 102% since purchase.

Three names carry most of the blame:

• CCC Intelligent Solutions (CCC, −33%) and Constellation Software (CSU, −32%) remain caught in the SaaSpocalypse, the market’s conviction that AI will let anyone rebuild established software.

• Copart (CPRT, −25%) is deeply out of favor, mostly driven by slowing revenue growth over the last several quarters. The shares are now down more than 50% from their peak, the largest drawdown in over 20 years. But as I argued in my recent Warning Signs post, the fundamentals seem to tell a different story. I still believe Copart is a wide-moat compounder working through cyclical challenges.

The Guru Gems portfolio performance stands in sharp contrast to the more ‘mechanical’ approach of my Magic Formula portfolio, which is up 16% since inception and 12% YTD, four percentage points ahead of the S&P 500’s 8%.

With high quality stocks underperformance currently at 1999 extremes (there is an interesting chart on this in Wedgewood’s latest letter; more on that further down this post), it’s too early to throw in the towel and give up on my Guru Gems portfolio.

So keeping François Rochon and Bryan Lawrence’s ‘rule of 3’ in mind, I’m giving myself another 2-3 years to see if I can do better.

🔍 Alphabet: A Phenomenal, Historic, Confusing Quarter

Alphabet released Q2 earnings this week, and since it is still my largest position, it deserves a closer look.

On the surface, the numbers look phenomenal:

The good:

• Revenue grew 24% to $119.8B: remarkable acceleration for a company this size. Search & other grew 17%, YouTube ads 13%.

• Google Cloud grew 82% (!!) to $24.8B, with operating margin jumping from 20.7% to 35.6% and backlog reaching $514 billion.

• Operating income grew 30%, faster than revenue; Alphabet keeps getting more profitable while investing heavily.

• The AI flywheel is spinning: the Gemini app reached 950 million monthly active users, and Alphabet’s model APIs now process ~22 billion tokens per minute, up from 16 billion just last quarter.

The bad (or at least, the expensive):

• Capital expenditures hit $44.9 billion in a single quarter, double of last year’s Q2, and management raised full-year capex guidance to $195–205 billion, with 2027 expected to be significantly higher again.

• As a consequence, for the first time in its history as a public company, Alphabet reported negative free cash flow: −$5.9 billion for the quarter.

• To fund its massive AI infrastructure push, Alphabet raised $49.6 billion of equity in June: a $20.5B public offering of common stock, a $10B private placement to Berkshire Hathaway, and $19.1B of 6.25% mandatory convertible preferred stock. This is on top of ~$56B of new debt in the first half. It also put a $40B at-the-market program in place (unused so far, and earmarked primarily to cover taxes on employee stock grants). Meanwhile, share buybacks have stopped entirely: zero in H1 2026, versus $28 billion in H1 2025.

• One more thing to keep in mind when seeing headlines about net income tripling: reported EPS of $9.11 (+294%) was inflated by a $99 billion mark-up of Alphabet’s equity investments, mostly its SpaceX stake. Strip that out, and the operating story is (only!) a 30% profit grower.

Alphabet is probably one of the most covered stocks on the planet, so rather than pretending I can add much new analysis, let me share some of the more interesting perspectives I came across this week:

• A good summary of the Q2 results by : Alphabet Inc. Earnings Update

• A very comprehensive deep dive on Alphabet by and how Alphabet is the only company paid at every layer of the token economy

• A much more bearish view by who expects Alphabet shares to fall much further than they already have this week

• One element that seems underreported: Alphabet’s equity issuance, and how it may be contributing to the current negative sentiment. Here is an interesting chart from Jeff Weniger

  

• Finally, an interesting take arguing Google’s search revenue growth is partly artificial and damaging for Google long term: Max Anderson on X

The market’s verdict, for what it’s worth: the stock fell about 7% the day after earnings and nearly erased all of its YTD returns. Over the past year the stock is still up 64% though…

As I shared in my Q1 13F update post, I trimmed my Alphabet position a bit in May at $395. Other than this small trim, I am keeping the position unchanged for now.

Subscribe now

→ If you enjoy this newsletter, please subscribe and like this post 🤍

→ If you find value here, consider buying me a virtual coffee ☕

⚕️ A Quick Word on Baxter (BAX)

Last week I re-examined the Baxter thesis and landed on "holding, not adding".

The next test comes quickly: Baxter reports Q2 earnings this Thursday, July 30. I'll be watching the three indicators I laid out last week: operating-margin progress toward the 13–14% guide, any timeline on lifting the Novum pump hold, and progress toward the 3.0x leverage target.

I’ll share a Note with my thoughts on Thursday or Friday, so make sure to subscribe/follow on Substack if you haven’t already (for all email-only subscribers, consider downloading the Substack app as this is where I share additional notes or interesting articles).

💸 A Quick Word on PayPal (PYPL)

Until last week, PayPal was also among the worst performers in the portfolio. That changed on July 15 when PayPal received a joint $53 billion takeover proposal ($60.50 per share) from Stripe and Advent International. The shares jumped ~17% on the news and have come back closer to my average purchase price of $59.9.

PayPal’s board of directors rejected the offer, signalling that the offer significantly undervalues the company.

I am not selling (yet) and will be watching closely how this further develops. PayPal is also releasing Q2 earnings this coming week, so it will be interesting to hear management’s outlook and any new comments on the takeover interest.

📜 Wedgewood Partners: Are Hyperscalers Still Magnificent?

The final piece this week: Wedgewood Partners’ Q2 letter, which connects nicely to the Alphabet story above.

Wedgewood Partners Q2 2026 Client Letter

3.88MB ∙ PDF file

Download

Download

I always enjoy reading Wedgewood’s Client Letters and I highlighted them as one of the picks in my 3 Favorite Investor Letters back in January.

Magnificent

Their latest letter contains a section titled “Are Hyperscalers Still Magnificent?”

Their answer: “We certainly think so, and have increased our weightings accordingly.”

A few interesting takeaways from the letter:

1/ The market is funding the AI trade by selling the hyperscalers

Wedgewood identifies three headwinds: negative sentiment around capex crushing profitability, billions rotating out of hyperscalers into semiconductor and memory stocks (semis are now ~25% of the S&P 500, up from ~5% a few years ago), and IPO funding for SpaceX (and prospectively Anthropic and OpenAI) coming at the expense of hyperscaler shares.

2/ Reframing the capex debate

Rather than staring only at collapsing free cash flow (see Alphabet above), Wedgewood looks at returns on the growing asset base:

“We think Alphabet is an exceedingly rare, if not entirely unique, business, growing a nearly $450 billion gross asset base by 40% while maintaining a 30% return on that massive asset base.”

They also make a point I haven’t seen anywhere else: the hyperscalers’ investment portfolios act as a hedge on their own capex inflation. Alphabet’s ~$150 billion of (unrealized) gains on SpaceX and its 14% Anthropic stake could effectively cover years of incremental memory-driven capex costs. Meta’s AMD warrants (~$90B by Wedgewood’s estimate) play the same role. The market ignores these “one-time” gains; Wedgewood argues they are an overlooked shock absorber.

3/ It's the second derivative that matters

Wedgewood has owned semiconductor stocks for decades (”we have the scars to prove it”), and their explanation of how cyclical top traps are set is worth reading:

“Memory company earnings will surely grow over the next few years, at least until demand cools and/or supply shortages wane. However, and this is key, it only takes a modest cooling in current red-hot demand or a modest easing of the significant supply shortage for these stocks to drop as suddenly as they have risen because expectations reverse; earnings expectations will always be too high once growth-rate deceleration kicks in. It is the second derivative change in the rate of growth that matters. This is how cyclical top traps are set. The market always sniffs out a peak in earnings growth acceleration well before the cycle turns. Again, it matters little if earnings continue to grow; the stocks lead fundamental results, often by years.”

If that sounds familiar, it’s essentially the same message Rich Pzena delivered in last week’s issue: today’s memory earnings are a cyclical peak, not a plateau.

4/ And finally: Zoetis, not so magnificent anymore

Wedgewood sold their entire Zoetis position this quarter. There is no explanation why and this seems like a strange move, as not long ago, they highlighted Zoetis as a high-quality business finally on discount:

“We believe that, over the long term, reality tends to win; high-quality businesses earn the valuations they deserve at some point and will outperform the broad market. We believe we have been given the opportunity to do this with Zoetis now.”

That makes two disciplined quality investors (Terry Smith at Fundsmith and David Rolfe at Wedgewood) who built Zoetis positions in early 2025 and have now walked away. (Smith did offer an explanation in his half-year letter: he pointed to management’s failure to respond to new generic competition, or to communicate clearly about it.)

Not everyone is heading for the exit, though. Michael Burry started buying in May, after the stock dropped further on a guidance cut. He believes the company is very well run, with strong capital allocation and conservative financing, in his words, a “fat pitch” that requires patience.

I made similar points in my May update when I added Zoetis to the portfolio, in addition to the exposure it gives to a powerful secular tailwind: the growing global animal health market.

That thesis hasn’t changed since May, so I’m happy to keep holding.

That’s it for this week…thanks for reading and I’ll be back next week.

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

In today’s update

• 🔭 Pzena’s Mid-year outlook for Value

• 📈 Deep Discounts in Deep Value

• ⚕️ Baxter - A closer look at one of my Guru Gems holding

New to Guru Gems? Start here.

Here is the 2 minute AI audio version of today’s post:

🔭 Pzena’s Mid-year Outlook for Value

I always enjoy listening to Rich Pzena. Pzena was Joel Greenblatt’s Wharton classmate and co-author of their 1979 research project that validated Ben Graham’s stock picking formula. Greenblatt once called him “one of the highest quality, smartest, nicest people I know”.

Earlier this week, Pzena Investment Management shared their 2026 mid-year outlook webinar.

Rich Pzena, together with portfolio manager Akhil Subramanian, shared what really matters for value investors in this volatile year. Three things stood out to me.

1/ The third technology cycle

Rich Pzena has now lived through three technology cycles as a professional investor: the internet bubble, the fintech boom, and now AI. His observation is that they all rhyme in the same specific way:

“The market sees it as great for the new entrants and poor for the incumbents... It’s not new technology versus old technology. It’s the purveyors of the technology versus the customers of the technology.”

In 1999, you could buy virtually any Western manufacturer for about five times its cyclical earnings power, while the internet builders were priced as if we already knew who would win. Instead, virtually every dominant incumbent adopted the new technology, ended up with higher margins than before, and the stocks were “doubles and triples and quadruples”.

Same story with fintech. Rich recalls the Citigroup CEO looking back in 2017 at the endless stream of 28-year-old founders who had promised to put him out of business five years earlier:

“Now I have a new crop of 28-year-old CEOs of startup fintech companies who are coming to sell me their technology, because they understand they don’t want to be a bank.”

The lesson for the AI era: don’t assume the customers of the technology are the losers. History suggests the opposite.

2/ Memory: peak, or plateau?

The most extreme exuberance today is in memory chips. Pzena’s view:

“While AI will change the world, it isn’t going to negate the cyclicality of memory demand.”

Akhil Subramanian on that cyclicality:

“My 8-year-old has lived through three memory cycles. He was born, there was a peak, then a trough, then a peak, then a trough, and now we’re at a peak again. My 5-year-old has lived through a cycle as well.”

Three years ago, Micron’s gross margins were roughly break-even. Today it is guiding to 85% gross margins, for an asset-heavy manufacturing business, higher than most asset-light software companies.

What’s even more striking: Micron, Samsung and SK Hynix currently earn higher gross margins than Nvidia and TSMC, the companies everyone considers the widest-moat monopolies in the industry.

Many argue memory has reached a “new plateau”. Pzena Investment instead anchors on normalized, mid-cycle earnings power five years out, factoring in new supply, potential Chinese entrants, and demand destruction from rising device prices:

“So we’re living in a moment where the demand is insatiable and everyone believes that this might be a plateau and we may not go back to the past cyclicality.

What we try to anchor ourselves to is what is the normalized earnings power of this industry based on competitive dynamics. So when we say normalized earnings power, we’re talking about 5 years out on a midcycle basis..

On the basis of normalized earnings, it’s not entirely obvious to us that this is a new plateau... it’s probably fair to say the earnings are at a peak.”

The lesson here is something I already highlighted in my Pzena deep-dive: value on today’s earnings is not value if today’s earnings are a cyclical peak.

3/ Opportunities in software

While the market crowds into the “purveyors” of AI, Pzena has been buying the customers. Specifically, software companies crushed by the SaaSpocalypse: the fear that AI lets anyone replicate established software over a weekend.

In the webinar they discuss 2 names:

Workday (WDAY) — mission-critical ERP software for people-centric businesses (think Chipotle or Walmart, hiring thousands of people a week). Retention runs at 97–98%, and AI cuts both ways in their favor: it powers new products (a conversational AI tool that filters 2,000 job applications for a single Chipotle store) and reduces R&D, the biggest cost bucket of any software company. Pzena calls it “a very compelling investment opportunity.”

SAP — another ERP giant, serving supply-chain-heavy customers like BMW and Mercedes. Same mission-critical profile, plus a catalyst: as SAP retires its decades-old ECC product, most of that installed base is migrating to SAP’s next generation, which is a multi-year revenue uplift the market is undervaluing.

If you read last week’s Guru Gems issue, the following description of WDAY and SAP will sound familiar:

Mission-critical system of record. Retention near 100%. Enormous switching costs. AI as an accelerant rather than a death sentence. Sold off indiscriminately in the AI panic.

That is, almost word for word, the Guidewire (GWRE) thesis from last week. Guidewire’s ~99% retention is even higher than Workday’s.

Two Gurus with completely different styles (Lawrence the concentrated quality investor, Pzena the deep-value contrarian) are finding opportunities in the same corner of the market.

→ If you enjoy this newsletter, please subscribe and help others discover it by clicking 🤍

→ If you find value here, consider buying me a virtual coffee ☕

Subscribe now

📈 Deep Discounts in Deep Value

I came across a piece from Oldfield Partners — Away From the Casino Tables: The Wildest Value Gap Since 1999 — which connects beautifully with Pzena’s webinar.

It was the most widely read piece in Latticework‘s Weekly Inspiration newsletter last week. I highly recommend subscribing; there are always fantastic articles.

The core of the piece is one chart and one statistic. Splitting global equities into four valuation buckets and comparing each to its own 40-year median:

• Deep value globally trades at a 45% discount to its historical average

• Extreme growth trades at a ~40% premium

The author, Sam Ziff, notes that only once before (late 1999, early 2000) has the market been more expensive than today based on the Shiller P/E ratio. Warren Buffett recently observed that “we’ve never had people in a more gambling mood than now”.

The last time markets were this concentrated and this expensive, the following decade (2000–2009) delivered roughly zero total return for the S&P 500 and MSCI World, while emerging markets doubled and value beat growth. This should serve as an important warning to index investors.

Now connecting this to what Rich Pzena said in the webinar. Asked whether value investing is broken after another stretch of index outperformance:

“Value investing can’t be broken... if you do it right and buy things for less than they are worth, you will win.”

And on today’s valuation spreads:

“Our portfolio sells at about 20 or 30% below the valuation [forward P/E] it did 10 years ago... and the market is roughly 30% higher than it was 10 years ago on the same metric. The spreads are as wide as they get.”

Betting that the same expansion repeats from today’s starting point is, in Pzena’s words, “a tough bet”.

Or as Oldfield puts it:

“The casino may stay open a while longer. However, the best returns over the next decade are unlikely to come from the tables everyone is crowded around.”

The point is not to see this as a signal to time the market, but rather as a signal about where the next decade’s returns are likely to come from.

Which brings me to one of the most out-of-favor stocks in my portfolio…

💎 A Gem Revisited: Baxter (BAX) — Hold, Add, or Fold?

The setup: Baxter is a 95-year-old medical technology company selling essential hospital products: sterile IV solutions and infusion pumps, smart hospital beds and patient monitoring, and specialty injectables. Its North Cove plant produces roughly 60% of all IV solutions used in US hospitals.

Baxter is deeply out of favor by any objective measure: the shares have lost over 70% over five years while the market sits near record highs, Wall Street's consensus rating is Hold with an average price target below the current share price, and the dividend cut expelled the income-investor base.

Exactly the kind of situation Rich Pzena hunts for. At the end of Q1, Baxter was among Pzena’s top-5 positions, and at the end of Q1 Pzena Investment Management owned more than 10% of the entire company.

In the mid-year webinar, when asked where he sees value today, Pzena pointed to healthcare, and named Baxter specifically:

“You have companies like Baxter, which hasn’t really rallied and has a very strong set of businesses that have earned high margins, strong free cash flows, modest growth rates — that have run into problem after problem after problem that the former management team was unable to solve, and the new management team is in the process of solving...

It’s priced as if they’re not going to succeed. And so if they do, you make a lot of money. If they don’t, you don’t lose a lot of money.”

I first bought Baxter in September 2025 around $23, added below $20 in November, and again around $20 in January. My average purchase price is $20.4. The stock closed at $22.6 this week, putting my position up 10.6%, and making Baxter a 5–6% position in the Guru Gems portfolio. A good moment to re-examine the thesis.

1/ What has gone wrong (a lot)

First a reality check. Since my original write-up, the “problem after problem after problem” list has grown:

• The Novum IQ pump hold. In July 2025, Baxter voluntarily paused US/Canada shipments of its Novum IQ infusion pump after reports of underinfusion were linked to 79 serious injuries and two deaths. The stock fell ~21% in a day. The hold remains in place, ceding pump share to BD and ICU Medical, and has triggered a securities class action.

• The dividend is essentially gone. In late 2025, Baxter cut its quarterly dividend from $0.17 to $0.01 (a 94% cut, ending a 55-year streak) to free up >$300M per year for debt reduction.

• The balance sheet is stretched. ~$8.6B of long-term debt remains from the 2021 Hillrom acquisition. S&P and Moody’s both downgraded Baxter in November 2025 to one notch above junk, and the 3.0x net-leverage target slipped from end-2025 to end-2026.

• Margins and earnings are going the wrong way. Adjusted gross margin fell 900 basis points year-over-year in Q4 2025, adjusted operating margin is guided at just 13–14% for 2026, and adjusted EPS is guided down to $1.85–$2.05 from $2.27 in 2025.

• Management churn. CEO Joe Almeida left abruptly in early 2025; new CEO Andrew Hider (ex-ATS and ex-Danaher) only started in August 2025, and the CFO seat is currently held on an interim basis.

This is not a quality compounder having a bad quarter. This is a genuinely wounded business. So why hold it?

2/ Is it ‘Good’?

The core franchise survived all of the above intact:

• Irreplaceable infrastructure: ~60% of US IV fluids from one vertically-integrated plant. When Hurricane Helene knocked North Cove offline in 2024, US hospitals rationed IV bags, which shows how few substitutes exist.

• Switching costs: infusion pumps and smart beds are wired into hospital electronic medical record (EMR) systems, drug libraries and biomed workflows

• Essential demand: hospitals buy these products in every economic scenario.

Morningstar still assigns a narrow moat, though it flags real erosion at the edges: pump share losses during the Novum hold, IV share slippage during the Helene outage, and generic price pressure in injectables.

Verdict: an essential, moaty core wrapped in a badly managed decade.

3/ Is it ‘Cheap’?

• At $22.6, Baxter trades around 11–12x forward earnings, which is a steep discount to the medtech peer group (typically mid-teens) and to its own pre-2022 history.

• Analysts’ average price target is ~$21.50 (consensus: Hold), which is below the current price. The market is pricing in that nothing improves.

• If new management delivers the margin recovery and the Novum hold lifts, earnings could reach ~$2.40+ by 2027 and a re-rating toward 13–14x would imply a stock around $30. If execution slips further, trough multiples on ~$1.80 of earnings suggest ~$15.

From $22.6, that is roughly +33% in the good scenario versus −34% in the bad one. This ‘bet’ only remains attractive if you believe, as Pzena does, that the odds of the fix are meaningfully better than a coin flip: the problems are operational rather than structural, demand is not in question, and a cost-focused outsider CEO with 92% performance-linked pay now runs the company.

Verdict: statistically cheap, but cheap for a reason. The bet is on execution, not on the multiple.

4/ To Hold or Not to Hold

Last week I scored Guidewire against Lawrence’s six questions. Let me do the same for Baxter and score it against Pzena’s own framework from my original deep-dive:

1. Is the business under-earning versus its historic norms? ✅ Operating margins of 13–14% sit well below medtech peers, returns on capital are currently below Baxter's cost of capital, and earnings are depressed by fixable operational issues (pump hold, hurricane costs, stranded costs, mix), not by falling demand.

2. Are the problems temporary and fixable, not structural? ⚠️ ✅ Pump holds, hurricane damage, bloated costs and bad M&A are all fixable in principle, but this management team hasn’t proven it can fix them yet.

3. Is the market pricing the downside case? ✅ ~11x forward earnings, consensus price targets below the share price, a departed income-investor base, and a 70%+ drawdown from the 2021 highs.

4. If they fix it, is the outcome skewed in my favor? ⚠️ Only with better-than-even odds on execution, and less skewed at $22.6 than it was at my $20.4 average cost.

📌 Final decision: Holding, not adding

Baxter stays in the Guru Gems portfolio as my designated deep-value turnaround, but I’m not adding at ~$23.

Baxter reports Q2 2026 earnings on July 30, its first quarter under a newly announced two-segment structure. Three things I’ll be watching:

1. Operating-margin progress toward the 13–14% guide;

2. Any timeline on lifting the Novum pump hold. Current guidance assumes the hold stays in place for the entire year, so any lift would be pure upside versus guidance;

3. Progress toward the 3.0x leverage target.

If those indicators start moving the right way, Baxter could be the kind of skewed outcome Pzena describes. If margins stall again, I may need to consider if holding this company still makes sense.

That’s it for this week! Please like or share if you enjoyed this post.

Thank you for reading and following along.

Until next week!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

For those of you that have been following Guru Gems for a while, you know that I am a big fan of the Richer Wiser Happier (RWH) podcast. A few weeks ago, William Green reshared clips from his interview with Bryan Lawrence, and the timing couldn't be better, since it's also the one-year anniversary of my deep-dive on Bryan Lawrence and Oakcliff Capital.

So this week, I revisit Bryan Lawrence one year later: what he said, what he’s been buying, and what he’s building.

I’m also taking a closer look at one of Lawrence’s holdings: Guidewire Software (GWRE)

Here is a 2-minute AI audio version of today’s post:

Guru in the spotlight: Bryan Lawrence

There is a lot to unpack in the below 30-minute clip from William Green’s interview with Bryan Lawrence. It is one of the best 30 minutes you can spend if you want to understand what a real investing process looks like: how ideas get sourced, filtered, researched, sized, and (inevitably) how some of them go wrong.

1. 🎲 The Rule of 3

A recurring theme in the clip is how Lawrence expects to be wrong about 30% of the time.

“If you measure being right as the stock that we bought, either we sold for more than we paid for it, or we own it still for more than we paid for it — we’ve been right 70% of the time and wrong 30%. Like, we’re definitively wrong 30% of the time.”

The last of Oakcliff’s six research questions (more on those below) also builds directly on this:

“Question number six: if we’re wrong — because we can expect to be wrong at least 30% of the time — how much money will we lose?”

Lawrence also points to Buffett’s famous 1984 “Superinvestors of Graham-and-Doddsville” essay: every single one of the Graham-school managers with market-crushing long-term records underperformed roughly a third of the years when measured annually. Lawrence went back and counted for Berkshire itself: 19 of 54 years (i.e. 35% of the time) Berkshire Hathaway stock underperformed the S&P 500.

If this number sounds familiar to Guru Gems readers, it should.

It’s part of François Rochon’s ‘Rule of 3’, which I covered in my Rochon deep-dive:

“One year out of three, the market will go down (by more than 10%). One stock out of three that we will buy will be somewhat disappointing. And one year out of three, we will underperform the indexes.”

Two great investors converging on the same ‘law’ for concentrated stock picking.

2. 😣 The Pain of Underperformance

Knowing you’ll be wrong 30% of the time is one thing. Living through it is another.

Green counted that Lawrence underperformed the S&P in 8 of his 20 years, and still beat the index by more than 100 percentage points cumulatively. Lawrence’s response to how he endures those stretches is really a masterclass in temperament.

An interesting data point he shares comes from DALBAR’s studies of mutual fund flows:

The average active manager underperforms the market by roughly their fees (~1% per year). Painful, but survivable. The real damage is how the clients behave: they add money after strong performance and pull money after weak performance. The result, per DALBAR:

“The client’s returns underperform the active manager’s returns by four percentage points… Four percentage points of underperformance over a 30-year investment horizon leading to someone’s retirement is a nest egg that is 70% less.”

This is quite a statistic. Investors don’t just underperform the market, they underperform their own funds, by a ruinous margin, purely through timing their entries and exits emotionally.

Lawrence adds a nice thought experiment:

“It’s 1975, over the prior 2 years Berkshire Hathaway stock has fallen by half and the market is flat. Do you sell your stock? If you had, that would have been a very bad decision.

It’s 1999, Berkshire Hathaway stock is up slightly but 40 percentage points less than the S&P. Do you sell your stock? That would have been a very bad decision in 1999.”

It’s clear that both would have been really bad decisions. Berkshire has lost half its value three times under Buffett — and underperformed in a third of all years — on its way to one of the greatest compounding record in history.

Which brings me to Terry Smith.

Terry Smith just released his Semi-Annual Letter to shareholders, and it contains the biggest changes in Fundsmith’s 15-year history:

• 13 names out: Atlas Copco, Coloplast, EssilorLuxottica, Intuit, LVMH, Nike, Novo Nordisk, Zoetis, Unilever, Mettler-Toledo, Otis, Wolters Kluwer, Magnum Ice Cream

• 12 names in: AppLovin, Netflix, Mastercard, Uber, Veeva Systems, TSMC, GE Vernova, Sage, Legrand, Nextpower, TJX, Yum! Brands

• Portfolio turnover in H1 2026: 51%. For a fund whose third commandment is literally “Do nothing”.

The fund was down 2.9% in H1 2026 while the MSCI World rose 11.2%, a 14-point gap in six months, on top of years of lagging a momentum- and AI-driven index.

Fundsmith has suffered massive redemptions, and it looks like this ‘Extreme Makeover’ is either an attempt to stop the bleeding or the result of significant pressure from the remaining investors to make drastic changes.

I have a lot of respect for Smith (he was the Guru of my very first issue!), but this level of change looks very strange and goes against many of Terry Smith’s own principles.

For further reading on Fundsmith’s Extreme Makeover, here are 3 posts I enjoyed:

• - Buy Good Companies, Don’t Overpay, Do EVERYTHING ...

• - Terry Smith’s Radical Strategy Shift

• - What Happens When Your Investment Philosophy Stops Working?

In closing, here is something which I realized is quite ironic: if you ask yourself who is actually living Smith’s famous mantra today...

1. Buy good companies ✅ Lawrence: natural monopolies, rational duopolies, low-cost operators + only businesses with durable cash flows

2. Don’t overpay ✅ Lawrence: his 15 holdings traded at ~14x trailing free cash flow at the time of the interview, vs. ~22x for the S&P 500

3. Do nothing ✅ Lawrence: roughly two new investments per year. “You may see us do nothing for a year.”

4. Strong performance ✅ Lawrence: while we don’t have the exact performance numbers, based on his reported US holdings, his performance over the past few years must have been very strong

...it looks a lot like Bryan Lawrence is applying Terry Smith’s mantra more than Terry Smith himself.

Subscribe now

If you enjoy this newsletter, please help others discover it by clicking 🤍

And if you would like to support my work, consider buying me a virtual coffee ☕

3. 🤖 Meet Marvin and Vicki

The second clip I want to share is Lawrence’s recent presentation at ValueX, and it’s a very different flavor: Lawrence on how he uses AI in the investment process.

Oakcliff has built two AI agents, running on Mac minis:

• Vicki is a shop-bot: she hunts for real-estate deals, cars, and restaurant reservations, partly to stress-test Oakcliff investments in areas where agents might disrupt the business model, partly to see what the technology can actually do. Verdict so far: not much. Zillow, Cars.com, OpenTable and friends have all blocked her. The internet has become, in Lawrence’s words, a game of “spy versus spy.”

• Marvin is an investment research analyst, and here Oakcliff is having far more success. Marvin applies Oakcliff’s frameworks, monitors the portfolio and watchlist, writes reports, and has a compounding memory base. Lawrence says Marvin can do “weeks of work in seven minutes.”

An interesting insight comes from Marvin himself, commenting on a draft of Lawrence’s presentation:

“My memory compounds correct and incorrect beliefs equally. Human discussion is the quality filter. Without discussion with you, I drift towards confident mediocrity.”

Lawrence’s conclusion: the quality of the agent’s work depends entirely on the quality of the frameworks you feed it. An agent will happily marinate in your biases and amplify them.

“What we’re trying to do with Marvin is have a small number of curated frameworks that maximize a query’s focus on information that is knowable, important, and true.”

There is one more gem in the Q&A that ties directly back to how Lawrence invests. Asked whether AI-driven quant trading changes the game, he lays out the market’s structure: roughly half is indexed, ~40% is algorithmic, and only ~1% is concentrated, long-horizon stock picking.

The algorithms are optimizing for what happens in the next ten days. But gradient descent (the mathematical engine that allows Large Language Models to "learn") doesn’t work when the outcome of a decision is only knowable in two or three years.

Lawrence’s edge (and the edge of everyone practicing this craft) lives precisely in that time horizon the machines can’t optimize for.

4. 📂 Oakcliff’s Q1 2026 Portfolio

So what does Oakcliff actually own, one year after my original deep-dive?

The answer is: pretty much the same portfolio as a year ago.

Here is the Q1 2026 13F snapshot (as of 31 March 2026, ~$227M in US-listed holdings):

Two things jump out:

1. Interactive Brokers (IBKR) is now a 30%+ position. IBKR was the Gem of my original Lawrence deep-dive, and it is also one of the larger positions in Marc Werres’s Hinde Group portfolio.

   With IBKR up ~200% over the past 2 years, this position has kept compounding in a true “let your winners run” allocation.

2. Lawrence has been aggressively adding to Guidewire Software (GWRE). He added 73% in Q1 2026 alone, on top of adding 27% in Q4 2025. He’s been buying while the stock has been cut nearly in half from its highs.

A Guru I highly respect backing up the truck on a wide-moat software company during an AI-panic selloff? That sounds like a Gem I need to take a closer look at…

💎 The Gem: Guidewire Software (GWRE) — To Buy or Not to Buy

The setup: Guidewire Software (GWRE) has been caught in the great Software Selloff of 2026. The market has decided AI will eat software, and anything with “SaaS” in its description has been repriced accordingly.

GWRE trades around $136, down ~27% year-to-date and roughly 50% below its 52-week high of ~$273. Lawrence first bought Guidewire in 2020 and has been adding significantly for two quarters.

Let’s run it through the framework.

1/ Is it ‘Good’?

Guidewire provides the core software (the ‘system of record’) for property & casualty (P&C) insurers: ClaimCenter (claims), PolicyCenter (policies), and BillingCenter (billing). This is not a productivity tool an insurer could turn off; it is the operational nervous system of the company.

In the RWH interview, Lawrence shares what defines a good business for him: natural monopoly, rational duopoly, or low-cost operator. Guidewire is the closest thing to a natural monopoly in P&C core systems:

• It is the leading provider to the industry, larger by revenue than its next several competitors combined, and wins more deals per year than its largest competitors combined (per Morningstar)

• Its software already covers more than 25% of all direct written premiums in the industry

• Customer retention is estimated at ~99%, making the duration of customer relationships longer than average within the software industry

• Switching costs are enormous: insurers keep core systems for two or three decades. Ripping out the system that processes your claims and policies risks the entire business

• In late 2025, Liberty Mutual signed a 10-year, Tier 1 deal that management called "one of the most strategic partnerships in our history"

Morningstar assigns a wide moat, and in March 2026, when Morningstar downgraded six wide-moat software companies (Adobe, Descartes, Manhattan Associates, Salesforce, ServiceNow, and Shopify) over AI disruption concerns, Guidewire kept its wide moat. Its analyst argues that a core system of record for a famously slow-moving, heavily regulated industry is relatively insulated.

Guidewire management also argues that AI is an accelerant: legacy systems from the 1950s can’t deliver AI capabilities, making modernization more urgent, not less.

Finally, the business is executing beautifully through the panic: fiscal Q3 2026 revenue grew 27% YoY to $373M with a 20.9% non-GAAP operating margin, both above the high end of guidance.

This looks like a great business. ✅

2/ Is it ‘Cheap’?

This is the more difficult part. Here are a few elements to help us determine whether GWRE is cheap:

• Versus its own history, GWRE has de-rated sharply: the stock fell from a 52-week high of ~$273 to the $120s–130s, compressing EV/ARR from roughly 18x toward ~9x, near multi-year lows despite accelerating ARR* growth.

• Morningstar’s fair value estimate for GWRE is $220 per share. At ~$136, the stock trades at a Price/Fair Value of ~0.6, which is one of the largest discounts in their software coverage.

• On a current free cash flow (FCF) basis however, Guidewire does not look cheap. GWRE is trading at 33 times FCF, which is on the higher end for a software business and a multiple investors usually pay for expected future cash expansion.

In other words: investors are expecting strong growth and expanding FCF margins driven by the cloud transition.

Verdict: cheap compared to its own history; reasonably priced on the margin-expansion thesis; not cheap on today’s cash flows. ⚠️ ✅

3/ To Buy or Not to Buy

Before I make my final decision, let’s score Guidewire against Lawrence’s own six questions:

1. Do we understand the business? ✅

   Guidewire provides cloud-based core systems and data/analytic solutions to the global property-casualty insurance industry, serving hundreds of insurers in 30+ countries in claims management, policy administration, and billing.

2. Is it a good business? ✅ ✅

   Near-monopoly, ~99% retention, decades-long switching costs, wide moat

3. Does management think like shareholders? ✅

   Consistent execution; Morningstar rates capital allocation Standard; Well thought out acquisitions that complement the portfolio

4. Is it cheap? ⚠️

   ~0.6x fair value, but the IRR depends on future margin expansion

5. Why is it cheap? ✅

   Indiscriminate AI-panic selloff vs. an insulated (possibly AI-advantaged) system of record

6. If we’re wrong, how much do we lose? ⚠️

   Hard to assess. If the AI disruption is real, the stock can go much lower still.

📌 Adding Guidewire Software (GWRE) to the Guru Gems portfolio

I plan to add a starter position (1-2%) to the Guru Gems portfolio in the next few days.

This is a good business undergoing a temporary crisis of confidence, with a Guru I highly respect significantly increasing his position.

Since I expect to be wrong 30% of the time, and if the AI bears turn out to be right about even core systems software, there is still real downside from here. So position sizing will be key and hence a small starter position.

🧩 One More Thing: Guidewire Meets CCC

Regular readers will remember that I already own CCC Intelligent Solutions (CCC) in the Guru Gems portfolio: another insurance software name held by several Gurus. So am I doubling up on the same bet?

No, and it could be helpful to get a basic understanding of the four layers of the insurance software stack.

Guidewire and CCC sit in different layers. Guidewire is the operational backbone an insurer uses to issue policies and manage the claim file; CCC is the specialized auto-claims network the claim gets routed through.

A single auto insurer commonly runs ClaimCenter as its system of record and pushes the physical-damage workflow through CCC. They don’t bid against each other.

So rather than doubling the same bet, I now own two dominant, ~99%-retention franchises in two different layers of the stack: the core (Guidewire) and the claims network (CCC). Both have been de-rated by the same indiscriminate software selloff. Same storm, different moats.

That’s it for this week’s edition!

Thank you for reading and following along.

Until next week!

✅ Further Reading

This week’s issue connects across the Guru Gems archive:

• Bryan Lawrence & IBKR: the original deep-dive, one year old this week.

• Timeless Lessons from François Rochon: the Rule of 3, the Podium of Errors, and why patience is the rarest quality.

• Cloning the World’s Best Investors: my very first Guru Gems issue. Here I introduced Terry Smith’s framework.

• Marc Werres & The Hinde Way: Marc Werres, the other guru with a massive IBKR conviction position.

• Physics for Investors #2: on feedback loops

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome to Round 10 of building my Magic Formula Portfolio

This one is a bit special: this week marks exactly one year since I bought the first 3 stocks on 7 July 2025.

Quick reminder if you are new to Guru Gems: while studying investing masters, I am creating two real-money portfolios to put my learnings into practice.

If you are new to Guru Gems, start here

Magic Formula Refresh

In ‘The Little Book that (Still) Beats the Market’, Joel Greenblatt suggests that a regular investor can beat the market averages by buying a group of ‘good’ companies at ‘bargain’ prices.

• ‘Good’ → High return on capital

• ‘Bargain’ → High earnings yield

Since he expects the magic formula to work on average, Greenblatt suggests to own a basket of 20-30 stocks, to buy this basket over the course of a year, and to hold each stock for exactly one year.

You can read more on Joel Greenblatt and his Magic Formula here.

You can also download my Joel Greenblatt e-book for free:

Joel Greenblatt e-book by Guru Gems

494KB ∙ PDF file

Download

Download

One Year In: Current Magic Portfolio Performance

Last month I added Booking Holdings (BKNG) in Round 9, bringing the portfolio to its full size of 25 stocks.

The overall performance of the Magic Portfolio after year one: +18%.

On a YTD basis, the Magic Formula Portfolio (+14%) is beating both the S&P 500 (+9%) and the NASDAQ Composite (+11%):

Some observations after 12 months:

• The big winners were not the obvious ones. Lantheus (LNTH, +110%), Catalyst Pharmaceuticals (CPRX, +57%), Consensus Cloud Solutions (CCSI, +55%) and EverQuote (EVER, +53%) lead the pack. None of these were names I would have picked on conviction so I’m glad the formula picked them.

• 19 of 25 positions are in the green. The formula does not need every stock to work. SIGA (-42%), H&R Block (-21%) and lululemon (-18%) are the biggest detractors. This is exactly what Greenblatt warns about: some cheap stocks are cheap for a reason, but that’s why you hold a basket of 25 names, allowing the formula to work on average.

“If you are able to stick with the magic formula strategy through good periods and bad, you will handily beat the market averages over time” — Joel Greenblatt

If you enjoy Guru Gems, please help others discover it by clicking 🤍

Subscribe now

Selling My First 3 Stocks

Greenblatt’s rule is to hold each stock for exactly one year, then sell. No exceptions and no second-guessing.

So this week, I am selling the 3 names I bought on 7 July 2025:

The first cohort ends with a losing average of roughly -6%: two modest winners and one heavy loser.

SIGA was my best performer in the early months after jumping nearly 40% on strong earnings, but it ends the year as my worst position. I held through the entire round trip, which was painful to watch, but part of the process.

What is interesting is that both metrics (ROCE and EV/EBIT) deteriorated, across all three names. None of the three would qualify for the formula today and this illustrates that the annual refresh is how the portfolio stays good and cheap, instead of drifting into “was good, no longer cheap”.

New Magic Formula Stocks

As I explain in detail in my first Magic Portfolio post, I use Joel Greenblatt’s Magic Formula screener combined with financial ratios data to identify the top stocks in terms of combined rank for Return on Capital (using ROCE) and Earnings Yield (using EV/EBIT).

This round, I extended the analysis with three additional criteria: P/FCF (an alternative valuation lens to EV/EBIT), Revenue 3yr CAGR (is the company still growing?) and Net Debt/EBITDA (how much debt is on the balance sheet?).

Two notes on this round’s selection:

1. I limited precious-metals miners to one. The raw screen currently ranks three gold/tin miners near the top (Alphamin, Andean Precious Metals, Wesdome). I guess this is a reflection of where we are in the metals cycle. Following the ranking blindly would have made this a mining round. I capped it at one miner and chose the highest-quality of the three: Wesdome.

2. I skipped HP Inc (HPQ) despite its high rank. HPQ ranked above two of my picks on the raw screen. But it has appeared on the Magic Formula screener for a very long time while its share price has declined over the past 5 years. You could argue that this is the classic profile of a value trap, and therefore I decided to pass.

Wesdome Gold Mines (WDO, Market cap $2.7B)

Magic Formula metrics: ROCE 63% (3yr: 42%); EV/EBIT 5.6x; Net cash

Wesdome Gold Mines is a Canadian gold producer with two high-grade underground mines: Eagle River in Ontario and Kiena in Québec.

What sets Wesdome apart from most miners is the balance sheet: completely debt-free, with the company buying back stock while funding growth internally. Revenue grew +55% over the past 3 years on the gold-price tailwind.

The Buckle (BKE, Market cap $2.1B)

Magic Formula metrics: ROCE 35% (3yr: 34%); EV/EBIT 8.1x; Net Debt/EBITDA 0.4x

A specialty apparel retailer from Kearney, Nebraska, known as a denim destination, operating ~440 stores across 42 states, including its own exclusive BKE brand.

This is the cleanest “boring quality” name on the screen: high returns on capital year after year, near-zero debt, and a consistent dividend.

Revenue has been relatively flat over 3 years though recent comparable-store sales are accelerating (+5.7% in Q1).

Cognizant Technology Solutions (CTSH, Market cap $20B)

Magic Formula metrics: ROCE 20% (3yr: 20%); EV/EBIT 5.6x; Net cash

Cognizant is one of the world’s largest IT services firms, operating across financial services, health sciences, products & resources, and communications/media/tech.

Cognizant is cheap for the same reason many other service forms have been hit: the market fears AI will commoditize what IT-services firms sell.

At a price below 10x earnings, with a net-cash balance sheet, revenue still growing, an expanding operating margin and $1.6B per year returned to shareholders via buybacks and dividends, the magic formula is suggesting that a lot of pessimism is already in the price.

What’s interesting is that Rich Pzena holds Cognizant as one of his largest individual holdings and further added to the position in Q1 2026. Here is an extract from Pzena’s US Focused Value Q1 Commentary:

Leading IT services provider Cognizant (CTSH) was weak, despite reporting robust 4Q earnings with peer-leading organic growth, due to fears around AI disruption to the services industry more broadly. We continue to see these fears as overblown and added to our position on weakness.

Pzena also still owns Baxter as one of his top 5 positions, a stock I added in September 2025 and which is up 11% since then.

This brings the portfolio to 25 stocks again, and ready for year two. From here on, every round will be both a sell and a buy: the Magic Formula flywheel is now turning.

That’s it for this week! Please like or share if you enjoy reading Guru Gems.

You can follow me on X @guru_gems and Substack @gurugems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

A few things on my mind this week:

• ⚠️ Warning Signs - Some red lights are flashing in the stock market

• 🛡️ Downside Protection - Legendary investor Seth Klarman on the one thing he cares about as much as upside

• 🚗 Salvage Cars - A fresh, honest look at one of my Guru Gems holdings that’s been bleeding: Copart

New to Guru Gems? Start here.

Last week I covered lesser-known Guru Dennis Hong, founder of ShawSpring Partners, who runs a concentrated portfolio of companies in the technology, internet, and consumer sectors

Here is the 2 minute AI audio version of today’s post if you are short on time:

⚠️ Warning Signs

🎶

“A warning sign
I missed the good part, then I realised
I started looking and the bubble burst”

Warning Sign - Song by Coldplay

I’m fairly sure Chris Martin had something other than equity markets in mind when he wrote these lines. But listening to this song earlier this week, I realized how well these first three lines capture the mood around AI stocks right now.

I missed the good part

The rally has convinced a lot of people that they've missed the ‘easy’ money and that the only way to fix that is to pile into the 'hot' names before it's too late.

Howard Marks described exactly this psychology in his book Mastering The Market Cycle:

“But most investors do capitulate eventually. They simply run out of the resolve needed to hold out. Once the asset has doubled or tripled in price on the way up — or halved on the way down — many people feel so stupid and wrong, and are so envious of those who’ve profited from the fad or side-stepped the decline, that they lose the will to resist further.

My favorite quote on this subject is from Charles Kindleberger: “There is nothing as disturbing to one’s well-being and judgment as to see a friend get rich” (Manias, Panics, and Crashes: A History of Financial Crises, 1989).

Market participants are pained by the money that others have made and they’ve missed out on, and they’re afraid the trend (and the pain) will continue further. They conclude that joining the herd will stop the pain, so they surrender. Eventually they buy the asset well into its rise or sell after it has fallen a great deal.

In other words, after failing to do the right thing in stage one, they compound the error by taking that action in stage three, when it has become the wrong thing to do. That’s capitulation. It’s a highly destructive aspect of investor behavior during cycles, and a great example of psychology-induced error at its worst.”

Look at what is happening in Taiwan and Korea and tell me this isn't exactly what Marks is describing…

I started looking and the bubble burst

Is there an AI bubble? Is it 1999 all over again? I’m seeing many posts on Substack and finance news sites debating these questions, claiming either that we are reliving 1999-2000 or that “this time is different”.

As with everything, the answer needs some nuance rather than bold statements.

I have written in a few previous posts about how history seems to be rhyming lately:

• What Rich Pzena learned from the dot-com bubble (Dec 2025)

• GMO: “It’s probably a bubble” (But that’s OK) (Dec 2025)

• Berkshire’s Underperformance - History rhymes (May 2026)

• Paul Tudor Jones on whether this is 1999 again (May 2026)

Rather than repeating myself by listing every warning sign I'm seeing today, let me share two excellent pieces I read this week that captured my current thinking:

• - Episode 33: This Is Not 1999, Mostly. (free)

• - “We’re In A Historic Bubble” (paid)
  

Come on in… I’ve gotta tell you what a state I’m in

So what does all this mean for my Guru Gems portfolio?

Starting with Alphabet and Amazon, my two best performers. Both have benefited from the AI rally, especially Alphabet which I bought in April 2025 when everyone thought they were done because Search was no longer relevant.

Both companies would almost certainly take a hit in a broad selloff (in fact, both are down 13% and 14% over the past month as part of a broader tech sell-off). But I see them as genuinely high-quality businesses that should keep compounding over the long run, so I'm comfortable holding them. They also remain two of the most widely owned names among the Gurus, as the latest 13F filings confirmed.

Then there is a group of names that have been sold off on the fear that AI will disrupt them (Constellation Software, CCC Intelligent Solutions, S&P Global). I probably bought CSU and CCC too early, but I remain convinced all three will do well over time, and the long-term Gurus’ activity backs up that thesis.

And finally, a few companies that are essentially AI-agnostic and have been left for roadkill despite still being high quality (Zoetis, Copart). These are exactly the kind of unglamorous, high-quality names that tend to do well once the euphoria around the hot names fades. When the dot-com bubble burst, these type of boring businesses quietly compounded while the hot names were collapsing.

→ If you enjoy this newsletter, please subscribe and help others discover it by clicking 🤍

→ If you find value here, consider buying me a virtual coffee ☕

Subscribe now

🛡️ Downside Protection

I watched 2 very interesting interviews with Seth Klarman this week. Klarman is a legendary investor who runs Baupost, the Boston-based firm he has led for almost 44 years. In those 44 years he has had only five down years, the worst around −10%. He is also the author of the famously hard-to-find book Margin of Safety.

I will be doing a full post on Seth Klarman in the coming weeks or months, so stay tuned for that one.

For now, I just want to share one insight that stuck with me from the interviews.

It is about the importance of downside protection.

Here is the line from the first interview that captures it:

It’s a downer to talk about downside. Nobody wants to hear it when the market’s going up every day. But there’ll be a moment when things will correct. There’ll be a lot of pain and usually in those environments if we’ve protected well which so far we have, we’ll be able to put capital to work when our competitors are maybe standing back a little bit. So the opportunities that happen in the worst environments can lead to returns for years afterward from deploying more capital into those markets.

And from the second interview:

“The only place where we’re making decisions is if the downside is protected and if we can see lots of paths to winning, then we’re very interested”

In almost 44 years, Baupost only had five down years. That record came from an obsession most of the industry finds boring. As Klarman put it, “the way we protect capital is we focus on downside as much as upside.” In practice that means meticulous bottom-up research, no leverage, a willingness to hold cash when nothing is compelling, and buying unloved or distressed assets during market panics.

Protecting the downside is precisely what lets you show up on the day everyone else can’t. Klarman’s example was 2008–09: as markets fell apart after Bear Stearns and Lehman, Baupost raised significant capital within a single quarter and deployed it into distressed assets while much of the industry sat “paralyzed and frozen with fear.”

There's a deeper, mathematical reason protecting the downside matters, which a fellow Substack writer I follow, of Rock & Turner, laid out beautifully in his article Are You Losing A Game You Should Win?.

Using a simple coin-flip (win +50%, lose −40% on a 50:50 flip), he shows why a strategy that looks favourable can still destroy capital: a loss always needs a larger percentage gain to recover, so the typical compounded outcome is negative even when the average looks positive. His conclusion is perfectly aligned with what we just heard from Klarman: "large drawdowns damage compounding more than large gains help it."

💎 Copart (CPRT) — Undervalued or value trap?

Copart (CPRT) runs online vehicle auctions, primarily for salvaged and total-loss cars. Insurance companies rely on Copart to efficiently dispose of vehicles, while buyers around the world compete for inventory through Copart’s digital platform.

Copart has compounded capital through multiple cycles without relying on leverage or hype.

Its strength can be summarized in 3 dimensions

• Network effects: More buyers → higher auction prices → attracts more sellers (insurance companies) → more inventory → attracts more buyers

• Long runway: Increasing vehicle complexity and repair costs means more cars are written off

• Strong balance sheet: Cash is piling up, offering opportunities for buybacks and/or acquisitions

I first wrote about Copart in my review of Peter Keefe and his Rockbridge Capital portfolio.

I added the stock to the Guru Gems portfolio in January 2026, when it was down nearly 40% from its 52-week high. I added more in February when the stock kept falling.

The stock is deeply out of favor, mostly driven by slowing revenue growth over the last several quarters, and shares are now down more than 50% from their peak, the largest drawdown in over 20 years.

But the fundamentals seem to suggest a different story:

• Revenue 5yr CAGR: 9.8%

• ROIC 5yr Avg: 31.6%

Why I think this is still a high-quality compounder

First, let’s start with was has gone wrong:

• Volumes are soft. Unit volume has declined for several quarters. More Americans underinsured or uninsured (fewer claims → fewer salvage cars), plus some share shifting between insurers. Fiscal 2026 revenue is set to dip slightly year-over-year, the first stumble in a long time.

• A competitor woke up. IAA, Copart’s longtime #2, was bought by RB Global in 2023 and appears to be competing harder.

• The market lost patience. A couple of earnings misses compressed the multiple hard. The forward P/E has fallen from the 30s into the high-teens.

But none of that means the moat is broken:

• Pricing power is intact. Even with volumes falling, average selling prices have kept rising (mid-single digits globally, high-single digits for US insurance auctions). A broken business doesn’t get to raise prices into declining volumes.

• The balance sheet is a fortress. Roughly $4.2 billion in cash against essentially no debt, with ROIC around 29%. Morningstar still rates the moat Wide.

• Management is buying back stock. After years of hoarding cash, Copart repurchased about $1.6 billion of shares over the past year (~6% of market cap), most of it last quarter, with the stock below analysts’ fair value. This is the corporate version of Klarman’s “playing offense when competitors are sidelined.”

So: undervalued, or value trap? A value trap keeps getting less valuable with margins eroding and the moat crumbling. That isn’t what these numbers describe. This looks more like a wide-moat compounder working through cyclical challenges. The bear case to take seriously is “volumes staying soft longer than expected and the stock sitting in the penalty box for a year or two”. This would be painful, but survivable.

So I’m holding, and at ~$30, with the moat intact and management buying back stock, I believe the risk/reward has improved.

Red flag to watch for: if average selling prices start falling alongside volumes, that could mean the moat is genuinely eroding, and this is when I would revisit straight away.

That’s it for this week! Please like or share if you find these insights valuable.

Thank you for reading and following along.

Until next week!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

In today’s edition:

• 🕴️ The Guru: Dennis Hong — the shopkeepers’ son who grew ShawSpring from $10m to over $1bn, running just 5–10 stocks

• 🌐 The Playbook: Ecosystem Control — Hong’s framework for defining quality

• 💎 The Gem: Coupang (CPNG) — a new Q1 2026 position, and a stock Hong has held before

2-minute AI audio version of today’s post:

Guru in the spotlight: Dennis Hong

Background

Dennis Hong grew up as the son of two shopkeepers in Ontario (“We basically were purveyors of cigarettes and lottery tickets”) which is where he first learned about business. Neither parent went to university, Hong was the first in his family to do so.

One of his applications (“one of my lottery tickets”) hit, and he got into Yale. To cover ~$4,000 a semester he searched the campus job board for the highest hourly rate and landed, almost by accident, at the Yale Investments Office, which was run by the legendary David Swensen.

At the time he didn’t even know who Swensen was. Sitting in on manager meetings with the likes of Seth Klarman and Chase Coleman, and reading partner letters in the filing room, became his real education.

From there he spent a couple of years at Matrix Capital, then in 2009 he became employee number two at Altimeter Capital (founded by Venture Capitalist Brad Gerstner), which he helped scale from $2m to $400m.

Launching ShawSpring

At 30 he enrolled at Harvard Business School intending to network his way into a fund launch, and then dropped out after his first year when a Boston family office offered to back him. ShawSpring Partners launched on 15 July 2014 with $10m.

ShawSpring’s assets grew from $10m to over $1.5bn in less than seven years. (It did significantly decrease again over the past years, more on that later.)

There is no IR team and no marketing department. Hong is, in his words, the IR, the portfolio manager, the marketer and the CEO. The ceiling on the firm’s size, he says, is simply his own bandwidth to maintain a real relationship with every client.

Subscribe now

If you enjoy this newsletter, please help others discover it by clicking 🤍

And if you would like to support my work, consider buying me a virtual coffee ☕

ShawSpring’s Playbook

ShawSpring’s Partner Letters are unfortunately not publicly available and there aren’t a lot of recent interviews, but there are some older interviews and letters that contain enough to study Hong’s thinking process and investment approach.

Here is a chat between Dennis Hong and Robert Vinall which I much enjoyed:

Hong’s firm runs on a single organising principle: maximise the return on time and effort spent. With a five-person team hunting for 5–10 stocks out of a universe of ~50,000, everything has to be “repeatable, replicable and scalable”.

Here are 4 pillars of Hong’s framework which I distilled from the different interviews or available letters

1. 🌐 Ecosystem Control: lock-in across every constituency

One idea sits at the centre of how Hong invests: Ecosystem Control

Think of every business as an ecosystem. The customer is the catalyst that brings cash in, and that cash then flows through all the other constituencies: suppliers, manufacturers, employees, management, investors…

The highest-quality businesses are the ones where every constituency is locked in so completely that it would be economically irrational for any of them to leave for a competitor ecosystem.

That dynamic is rare, and it’s the filter ShawSpring uses to sort ~50,000 companies down to a handful.

“If it doesn’t have this characteristic of ecosystem control, you just sort of drop it and go to the next one.” — Dennis Hong

Underpinning it is a definition of compounding borrowed from Peter Kaufman: “constant, continuous, dogged improvement over very long time frames”. Hong wants businesses that grow revenue, earnings and free cash flow with that kind of positive linearity regardless of the macro environment. Ecosystem Control is his test for which ones can.

2. 🎯 Concentration and a punishing hurdle

ShawSpring runs just 5–10 stocks in service of what Hong calls “Hall of Fame returns”

Over his ~40-year horizon he says he is aiming for a “two-handle” annualised return (20%+), with a very high bar for entry of roughly a 30% IRR (internal rate of return) over 3–5 years. (Whether he’s actually clearing that bar, we can’t verify, but it’s a useful way of how to think about a hurdle rate.)

As he tells his team, unless an idea could plausibly triple or quintuple over that window, it isn’t worth the time.

The firm runs as “one team, one portfolio, one P&L” (“there’s no ‘sleeves’ in our firm, everyone owns every position”), and to fight group-thinking, Hong appoints himself ‘chief pain in the a**’: once a new idea is about a third researched, his job is to try to kill it.

With ten strong names already in the portfolio, the bar to displace an existing name is meant to be very high.

3. 🛒 The ‘trigger-ready’ list: patience as process

ShawSpring keeps a shopping list of ~200 businesses that appear to have ecosystem control.

And within that shopping list they have a ‘trigger-ready list’, a subset of fully researched businesses which they really like, but where the only criteria missing is a good price.

When a sell-off drags a business down to Hong’s hurdle rate, the team can act immediately. His March-2020 purchase of Square, after a ~60% drop, is an example of this.

4. 🔭 Optionality: paying for the upside the market can’t see

If Ecosystem Control is how Hong defines quality, Optionality is how he thinks about growth the market is missing.

In a 2020 letter, Hong defined Optionality as “the expected value of a current or future investment that is unknowable or difficult for the market to discern.”

ShawSpring Partners Letter Nov 2020

675KB ∙ PDF file

Download

Download

He argues that the market prices the visible core business well, but struggles with the less knowable, so it tends to draw a straight line from today’s numbers and ignore what isn’t yet on the page.

Hong identifies four types of optionality:

• New business — leveraging the core’s infrastructure (Amazon → AWS; JD.com → JD Logistics).

• Product / category expansion — complements that deepen the value proposition (a marketplace adding groceries; software adding modules).

• Strategic shift — a related pivot to a better model (Netflix DVDs → streaming; Adobe → SaaS)

• Geographic expansion — replicating success in a new region (Starbucks into Asia; Sea’s Shopee across markets).

The ShawSpring Portfolio

Below is an overview of ShawSpring’s US equities portfolio at the end of Q1 2026.

This is a portfolio of high-growth technology and internet names: Okta, Zscaler, Braze,… businesses earlier in their potential compounding curve than a typical Guru Gems holding.

It’s also worth looking at the holdings history to get a sense of the turnover in the portfolio

Here is where I’m struggling a bit to match the frameworks and what I can actually observe. With no audited returns, the public 13F filing is the only view we have, and it may not be the full picture (US-listed long positions only, so no non-US names or cash).

But it’s still worth looking at it, because it raises some questions:

• The turnover looks relatively high for a framework built on the idea of ‘constant, continuous, dogged improvement’

  • Shift4 Payments (FOUR) was bought in Q1 2024, became the largest position by Q4 2024, then trimmed throughout 2025 (possibly taking profits as the stock peaked in early 2025), significantly adding again in Q3 2025, but then sold entirely during Q1 2026.

  • Intuit (INTU) became the largest position in early 2025 and is both much smaller and down sharply since.

  • Coupang (our Gem for today) was owned in 2022–24, exited, and is now back in the portfolio.

• The disclosed US equities portfolio has shrunk a lot. ShawSpring’s 13F value peaked over 1 billion in 2021 and stood at about $268m by Q1 2026. Some of that could be market drawdown, some could be client redemptions or it could be a move to cash or foreign equities (Dennis Hong did mention in an interview that he would be looking more at Asia). But a ~70% decline in disclosed US equities is still hard to reconcile with a smoothly compounding fund.

Of course, the above doesn’t prove the approach is flawed, a concentrated portfolio is volatile by design, and 13Fs certainly don’t tell the full story. But it does mean I can’t point to ShawSpring and say “look how well this works” like I may have done with Marc Werres and Hinde Group.

If anyone has access to ShawSpring’s performance or recent letters (@Dennis Hong, if you happen to read this ;-) ), I’d love to read them and will be more than happy to do a part 2 of today’s post.

💎 The Gem: Coupang (CPNG)

In Q1 2026, Hong re-opened a position in Coupang (NYSE: CPNG), the dominant South Korean e-commerce and logistics company often described as ‘the Amazon of Korea’.

It was one of four new buys in the quarter (alongside Zscaler, GDS Holdings and Constellation Energy), and he also added 57% to his Amazon position. With Alibaba as his 3rd largest position, his portfolio is clearly tilted toward scaled ecommerce platforms.

Hong is not the only Guru buying Coupang

I always look for Gems held by multiple Gurus and Coupang caught my eye in Hong’s portfolio because I had just read an investment thesis from Miller Value Partners on the same name. Here is the link:

Miller Value Funds - What We Own/Why We Own It - Coupang Inc. (CPNG)

Miller Value Partners was founded in 1999 as a joint venture between Bill Miller and Legg Mason Inc.

Bill Miller is most famous for his historic run managing the Legg Mason Capital Management Value Trust, where he beat the S&P 500 for a record 15 consecutive years (1991–2005). He’s definitely on my list of Gurus to cover in the future.

What does Coupang actually do?

Here is the introduction from Miller Value Partners’ write-up:

Dubbed the “Amazon” of Korea, Coupang is the leading e-commerce retail and logistics provider in the country. The company has over 100 fulfillment centers in Korea, with 70% of the nation living under 7 miles from a Coupang fulfillment center, enabling rapid delivery. In addition to selling owned inventory, Coupang sells advertising and goods from third-party merchants. It also provides restaurant ordering and delivery service through its Eats segment, while Coupang Play is a content-streaming service.

Coupang generates roughly $35bn in revenue across two segments: Product Commerce (~85% — its core retail, Rocket Delivery and WOW membership engine) and Developing Offerings (~15% — Coupang Eats, Play streaming, fintech, Taiwan and other bets). It serves Korea plus a growing international footprint, and employs over 100,000 people.

Not the first time Dennis Hong buys Coupang

Hong has owned Coupang before: it was the second largest ShawSpring holding in 2023 (around 22% of the 13F portfolio in Q2 2023). So it isn’t exactly a new idea for ShawSpring. While it’s unclear why they got out of the position in Q4 2024, I assume the reason they are getting back in, is because they believe it is a ‘great company temporarily on sale’:

• A late-2025 data-security incident and a large customer-voucher program hit margins, pushing Coupang to a Q1 2026 net loss of ~$266m on revenue of ~$8.5bn.

• The shares fell sharply, down roughly 30%+ year-to-date by spring 2026, resulting in multiple analyst downgrades.

• Underneath, the business kept growing (Developing Offerings +28%), membership recovered toward ~80% by April, and management expanded the buyback authorization to $2bn.

One of Hong’s favourite lines, from Intel’s Andy Grove, captures the bet he’s making on Coupang’s management:

“Bad companies are destroyed by crisis, good companies survive them, great companies are improved by them.” — Andy Grove

Why CPNG fits the Ecosystem-Control playbook

Coupang’s Rocket delivery network, WOW membership, Coupang Eats and Play streaming, and its third-party marketplace lock customers, merchants and the logistics fleet into a single web.

The more members join, the denser the delivery network; the denser the network, the harder it is for a customer or a merchant to leave. That is lock-in across constituencies, exactly what Hong screens for.

Let’s look at some of the key elements from Hong’s framework

🔒 1. Customers locked in ✅

Coupang has built one of the strongest consumer ecosystems in South Korea.

Rocket Delivery, Rocket Fresh, Eats, Travel, and the WOW membership create increasing switching costs. Once consumers become accustomed to same-day or next-day delivery and bundled services, changing platforms becomes inconvenient. The WOW membership is particularly powerful because it increases engagement across multiple services.

🏪 2. Suppliers/Merchants locked in ✅

Coupang is becoming an increasingly important sales channel for merchants, but merchants are generally less captive than consumers. Large brands can still sell through multiple channels, while smaller merchants benefit greatly from Coupang's logistics and customer reach. The marketplace is strong, but merchant lock-in is probably not yet at Amazon Marketplace or MercadoLibre levels.

🚚 3. Logistics moat (scale) ✅

This is probably Coupang's strongest advantage. The company has spent years building a nationwide fulfillment and last-mile delivery network that would be extraordinarily expensive to replicate. The density of its network creates faster delivery, lower costs, and better customer experiences, reinforcing its competitive position. Few competitors can justify the capital required to build a similar infrastructure.

🔭 4. Optionality ✅

Coupang scores exceptionally well here and fits all four of Hong’s types: new businesses, category expansion, strategic evolution, geographic reach.

The company continues to launch new businesses that leverage its existing customer base, logistics network, and brand. Advertising, Eats, Travel, Fintech, WOW, and future services all represent potential growth engines. Importantly, these initiatives are not standalone ventures, they are extensions of the existing ecosystem, which increases their probability of success.

To Buy or Not to Buy

Now that we’ve shown that Coupang fits the Ecosystem-Control playbook, let’s have a closer look at whether I should add it to the Guru Gems portfolio

Is it ‘good’?

As of 2026, Coupang’s gross margin sits at just below 30% and has nearly doubled over the past 5 years. This shows that Coupang’s warehouse automation, AI demand forecasting, and scale are working to lower product costs.

However, Coupang’s operating margin is razor-thin, hovering near 0-1%.

Management is choosing to reinvest its gross profits immediately into capturing new markets

• Building a massive delivery network (logistics moat)

• High expansion costs into new markets (Taiwan, luxury platform Farfetch)

• Heavy spending on its WOW membership ecosystem to maintain its dominant market share,

The recent data breach and massive $400m fine will also have an impact.

The investment thesis is that once Taiwan and other growth engines scale up and one-time legal issues fade, Coupang’s high gross profits should theoretically trickle down into a much stronger operating margin.

Is it ‘cheap’?

This is the part where I have done the least research on and therefore is a bit ‘too hard’ for me to assess right now.

When looking at some traditional metrics like EV/EBIT or P/FCF, Coupang may look very expensive.

However, Miller Value Partners writes that Coupang shares “trade at a forward (FY27) EV/Sales multiple of 0.6x, a ~71% discount to the stock’s peer group (AMZN, MELI, BABA, ETSY, EBAY, SE, Naver) average multiple”

Final decision

📌 Adding CPNG to the Guru Gems watchlist

Coupang ticks the Ecosystem and Optionality boxes, and the fact that it’s down nearly 40% over the past year means there is probably a lot of negative expectations already priced in.

But unlike a company like S&P Global, where the quality already converts to fat margins, Coupang’s profitability is still unproven.

So for now it goes on the Guru Gems watchlist rather than the portfolio. I do plan to study this business in more detail and will especially be looking for investment theses from other Gurus holding the stock (Oakmark, Chase Coleman).

That’s it for this week’s edition!

Thank you for reading and following along.

Until next week!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

Last month I launched a new series within Guru Gems: Physics for Investors.

In the first issue, I covered the second law of thermodynamics and the idea that businesses, like cups of hot coffee, naturally cool toward equilibrium unless work is done to hold them apart from it.

Today’s issue builds on that first one.

If the natural tendency of every business is to decay toward the cost of capital, then how do we explain the few businesses that seem to do the opposite? The ones that do not resist decay, but actually get stronger the bigger they get? Amazon, Visa, Alphabet…

These businesses are governed by a different principle which I will cover in more detail below.

2 minute AI audio version of today’s update:

New to Guru Gems? Start here

🎤 Why Microphones Scream

Picture a microphone held too close to a loudspeaker. A faint sound is picked up by the microphone, amplified, and pushed out through the speaker, now louder. That louder sound is picked up again, amplified, pushed out and so on… Within a second or two, the whole room is filled with a sharp, uncomfortable noise (a ‘screech’).

What you have heard is the result of a positive feedback loop.

Feedback is a fundamental concept in control theory and comes in two flavours:

Negative feedback is self-correcting. A classic example is a thermostat: the room gets too warm, the thermostat switches the heating off, the room cools back to the set point. Any deviation triggers a force that pushes the system back toward equilibrium.

As I explained in last month’s post, negative feedback is the reason most businesses drift back toward average returns. Competition is a negative feedback loop. It punishes success and erodes excess profit, always nudging the system back toward equilibrium.

Positive feedback is the opposite. Instead of counteracting a change, it amplifies it. A produces more of B, which produces more of A, which produces more of B again. The microphone screech is positive feedback. Or a snowball rolling downhill, gathering mass and speed as it goes. Each turn of the loop strengthens the next.

I will show later that the same concept applies to businesses: A negative feedback business fights just to stand still. A positive feedback business compounds.

One important thing to note is that positive feedback loops are almost always self-limiting. The microphone screech stops when the amplifier maxes out. The snowball stops at the bottom of the hill. Nothing compounds forever, because every loop eventually runs into a constraint: a saturated market, a physical limit, a competitor, a regulator…

⚙️ The First Feedback Machine

Last month’s Physics guru was James Clerk Maxwell. In 1868 Maxwell wrote a short paper titled “On Governors.”

A “governor” was the spinning-ball device that James Watt had attached to steam engines decades earlier to stop them running away with themselves: as the engine sped up, two weighted balls swung outward, closing a steam valve and slowing the engine back down. It’s the textbook example of negative feedback: a machine using its own deviation from a target to correct itself.

Maxwell was the first to analyze the mathematics of why such a device stays stable rather than oscillating wildly, and his paper is now regarded as the first significant paper on control theory. It was so dense and barely illustrated however, that it was almost entirely overlooked for eighty years.

The person who rediscovered it, and recognised its importance, was the guru of this month’s issue: Norbert Wiener. When Wiener built the science of feedback in 1948, he reached back to Maxwell’s forgotten paper as a foundation, and even borrowed his new field’s name from it. “Governor” comes, via Latin, from the Greek word kybernetes (χυβερνήτης), meaning “steersman.” Wiener took the same Greek root and coined the word cybernetics.

So the man who taught us about entropy also planted the seed of the science of feedback, which the man we are about to meet grew into a field.

Subscribe now

If you enjoy this newsletter, please help others discover it by clicking 🤍

🧔🏼‍♂️ The Guru: Norbert Wiener (1894-1964) - The Father of Feedback

Norbert Wiener was a child prodigy. He graduated from Tufts with a degree in mathematics at 14, briefly tried graduate study in zoology at Harvard (he gave it up after discovering he was hopeless at lab work), and completed a PhD at Harvard at 19, on mathematical logic.

He then studied in England under Bertrand Russell and in Germany under David Hilbert, two of the greatest mathematical minds of that time. In 1919 he joined MIT, where he would remain for the rest of his career.

Although Wiener is usually described as a mathematician, his most important early work sat on the border of mathematics and physics: Brownian motion, the random movement of tiny particles suspended in a fluid.

Brownian motion (sometimes referred to as a ‘Random Walk’ process) was first explained by Einstein in 1905, in one of the four ‘Annus Mirabilis’ papers he published that year). Einstein provided the first hard evidence that atoms and molecules actually exist.

15 years later, Wiener built on Einstein’s work and developed the mathematics that define the process. To this day, the process is called the Wiener process in his honour, and is denoted as W_t.

In 1948, Wiener wrote his most popular and influential book ‘Cybernetics: Or Control and Communication in the Animal and the Machine’ and coined the term cybernetics to describe the study of control and communication in machines and living beings. Wiener realized that “almost all complex systems are driven by feedback loops of information”.

For Wiener, a system becomes “intelligent” if it uses past outputs (feedback) to improve future behavior. His work showed that whether in a thermostat or the human brain, feedback loops allow a system to adjust its actions based on results.

Wiener’s book formally established how feedback stabilizes or destabilizes systems. As he noted, no one before had fully recognized how feedback and information processing keep systems stable. In doing so, Wiener bridged engineering and biology, influencing later fields like systems theory and AI.

📖 For more on Wiener's life and his line from cybernetics to modern AI, Max Planck Neuroscience has a short, accessible piece: From Cybernetics to AI: the pioneering work of Norbert Wiener. Britannica's biography is also excellent.

🌀 The Engines of Compounding

Wiener’s great insight was that you can understand a complex system by asking a simple question: what are its feedback loops, and which way do they run?

Applying that same question to a business can help us to identify compounders.

Most businesses are dominated by negative feedback. Grow too fast and you attract competitors. Get too big in one market and you saturate it. Raise prices too far and customers leave. These are all corrective forces (the competitive equivalent of Maxwell’s steam-engine ‘governor’), and they are why, as we saw last month, the typical business regresses toward the cost of capital over time.

Few businesses are dominated by positive feedback. For these companies, each unit of success makes the next one easier, not harder. Size becomes an advantage rather than a burden. These are the great compounders, and almost all of them run on one of four engines:

1/ Network effects

Each new user makes the product more valuable to every other user.

Visa is a textbook case: The more consumers carry Visa cards, the more merchants must accept them; the more merchants accept them, the more consumers want them. The network feeds itself.

Uber, which I wrote about through Bill Ackman’s lens, is based on the same idea: more drivers mean shorter wait times, which attract more riders, which in turn attract more drivers.

Mastercard and MercadoLibre, both on my watchlist, run on versions of the same loop.

2/ Scale economies

As a business grows, its cost per unit falls, letting it lower prices, which wins more volume, which lowers costs further.

This is the engine of Amazon and Costco for example. Wise, a cross-border payments company and a newer name I have started studying in more detail, also has this engine (more on Wise below).

The investors Nick Sleep and Qais Zakaria gave this model the memorable name “scale economies shared”. Jeff Bezos famously popularised the same idea as the “flywheel”, which is simply positive feedback drawn as a circle.

3/ Data and learning loops

More usage generates more data, which improves the product, which attracts more usage.

This is Google’s search advantage, and increasingly the engine of every serious AI business. The system gets smarter the more it is used.

4/ Brand and reputation flywheels

Success builds trust, trust builds more success.

Think of the credit-rating duopoly of S&P Global and Moody’s, or the slow-compounding prestige of a luxury house. Reputation, once earned at scale, becomes very hard for a newcomer to replicate.

These four engines nicely connect back to last month’s post. A positive feedback business is one whose compressor gets cheaper to run as the business grows. The network, the scale, the data, the brand: each does work of holding back competition with every passing year. That is why these businesses are so extraordinarily valuable, and so rare.

💳 WISE: A Scale Economies Case Study

Wise (formerly TransferWise) illustrates a powerful scale-driven feedback loop.

Wise built a global payments network enabling low-cost cross-border transfers.

The more money customers move through Wise, the lower its cost per transaction becomes, and rather than pocketing the savings, Wise passes them back as lower fees, which attracts still more volume.

This is a textbook positive feedback loop.

By FY2025 Wise was processing over $185 billion in annual transactions for about 15.6 million people and businesses.

I have a personal connection with Wise. I have been an expat for more than a decade and have used Wise for most of that time either to transfer money between international accounts, or to pay for things while travelling.

For years I was a happy customer without ever thinking of it as an investment. Only recently, when Wise came in the news quite a bit after shifting its primary listing from London to New York, did I start looking at Wise through an investor's eyes.

Here are some excellent deep-dives on Wise worth reading:

• - Wise: The Holy Grail Investment?

• - Wise: The Real-Time Dismantling of Cross-Border Banking

• - Deep Dive: Wise Plc. – Betting on the Unexpected?!

✔️ Three questions to ask before owning a compounder

In the spirit of last month's issue, here are three feedback questions worth asking before owning a business:

1. Which way does the feedback run? Does growth in this business invite correction (negative feedback) or does it reinforce itself (positive feedback)? Does getting bigger make the company stronger, or does it simply invite competitors and diseconomies of scale?

2. What is the engine, and is it still turning? If there is a positive feedback loop, name it precisely: network, scale, data, or brand. Then ask whether it is still strengthening or beginning to mature. Remember that every loop is self-limiting. A network effect in a saturated market is a spent force, however powerful it once was.

3. What could break the loop? A positive feedback loop can turn negative if external conditions change. Could new regulations, competitors, or technology shifts weaken the loop? For example, a company’s social network might flourish until a tough privacy law emerges, or a market leader might lose its edge if the industry “heats up” with insurmountable competition. Investors should check whether the business’s feedback mechanism still holds in tomorrow’s environment or if it’s at risk of reversal.

Closing Thought: The Investor’s Job

Negative feedback is everywhere. It is the governor of the whole economy, always pulling the exceptional back toward the average. Every so often, however, a business builds a loop that compounds instead of corrects, and those loops are where the big winners are made.

But finding the loop is only half the work. Every loop is self-limiting; the snowball always reaches the bottom of the hill.

The investor’s real job is not just to spot a positive feedback loop, but to judge how much room it has left to run.

That’s it for this week!

I hope you enjoy these ‘Physics for Investors’ issues. Please give me feedback (pun intended) on how to make these even more interesting or what topics you would like to see covered.

If you enjoyed this post, I would really appreciate it if you could like or share it or leave a comment!

You can follow me on X @guru_gems and Substack @gurugems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome to Round 9 of building my Magic Formula Portfolio.

Quick reminder if you are new to Guru Gems: while studying investing masters, I’m creating two real-money portfolios to put my learnings into practice.

Check out my recent 1-year anniversary post to learn more about my journey:

In the most recent Guru deep-dives I covered 2 lesser-known Gurus, who have both delivered exceptional results:

Magic Formula Refresh

In “The Little Book that (Still) Beats the Market”, Joel Greenblatt suggests that a regular investor can beat the market averages by buying a group of ‘good’ companies at ‘bargain’ prices.

• ‘Good’ → High return on capital

• ‘Bargain’ → High earnings yield

Since he expects the magic formula to work on average, Greenblatt suggests to own a basket of 20-30 stocks.

You can read more on Joel Greenblatt and his Magic Formula here.

You can also download my Joel Greenblatt e-book for free:

Joel Greenblatt e-book by Guru Gems

494KB ∙ PDF file

Download

Download

Current Magic Portfolio Performance

Last month I added lululemon athletica (LULU), Exelixis (EXEL), and MediaAlpha (MAX) to the portfolio in Round 8, bringing the total number of holdings to 24 stocks.

The overall performance of the Magic Portfolio is currently +8.2%, which is the best performance I have had until now.

Considering that many names are in small-cap, out-of-favor industries and that there is virtually zero exposure to semiconductor/AI stocks, an 8% return after a bit less than a year is not that bad.

• Lantheus Holdings (LNTH), a radiopharmaceutical company, is my best performer, and at some point was even a 2-bagger. The main driver was the FDA approval for a new formulation called Pylarify TruVu on March 6 2026.

• Catalyst Pharmaceuticals (CPRX) is my second-best performer with a 56% return since buying it last August. The big ‘catalyst’ here was the company securing extended U.S. patent exclusivity for Firdapse until 2035.

• lululemon (LULU), a name I added in the most recent round in April, sank more than 10% this week after reporting Q1 2026 earnings, in which they lowered full-year outlook. While this was somewhat expected, multiple analysts further adjusted their price targets downward, putting more pressure on the stock.

If you enjoy Guru Gems, please help others discover it by clicking 🤍

Subscribe now

New Magic Formula Stocks

As I explain in detail in my first Magic Portfolio post, I use Joel Greenblatt’s Magic Formula screener combined with financial ratios data to identify the top stocks in terms of combined rank for Return on Capital (using ROCE) and Earnings Yield (using EV/EBIT).

Since I am already at 24 names in the portfolio, I will be adding just 1 name in this round, to get to a full portfolio of 25 names.

#25 is a bit special since it is a ‘Magic Formula meets Guru Gems’ stock.

Booking Holdings (BKNG, Market cap $130B)

Magic Formula metrics: ROCE 3Yr: 72%; EV/EBIT: 13.6x

Booking Holdings is the world’s largest online travel agency, operating across five primary consumer-facing brands: Booking.com, Priceline, Agoda, KAYAK, and OpenTable.

The company serves travelers and accommodation providers in more than 220 countries, processing over 1.2 billion room nights in 2025, a new all-time record.

With $27B in revenue and $9B in free cash flow in 2025, this is a software-like business operating in a growing $3+ trillion global travel market.

Magic Formula meets Guru Gems

Not only does BKNG rank high on the Magic Formula screener (3yr ROCE 72%, EV/EBIT 13.6), but it also appears in the portfolios of some of the Gurus I am tracking.

François Rochon holds a ~3% position, AKO Capital has it at ~4% and for David Rolfe at Wedgewood Partners it is a ~5% position.

What’s happening

The stock is down more than 25% over the past year, now trading at around $165 (there was a 25-for-1 stock split in April).

Here is what David Rolfe wrote in his latest Q1 2026 Client Letter on why he believes the stock went down:

Booking Holdings detracted from overall performance during the quarter. Earnings per share grew +17%, with revenues up +16%, as travel demand remained strong late into 2025 and into early 2026. Most of the stock’s weakness stemmed from investors labeling it an “AI loser” and, later in the quarter, the outbreak of war in the Middle East.

Consumer AI, as a disruptive force in existing commerce, is proving to be much more difficult than markets expect. We view these AI tools and distribution channels as incremental rather than as drivers of the “zero-sum” dynamics that markets have been craving. As for the Middle East, for now, it represents a short-term disruption to travelers, particularly air travelers, given credible risks to international airline fuel supplies. However, we expect this will not be a multi-year headwind - more like a few quarters.

When both a quantitative screen like Greenblatt’s Magic Formula and long-term quality investors like Rochon converge on the same name, it’s usually worth paying attention.

On Monday, I’m adding BKNG to the Magic Formula Portfolio and I will then hold it for exactly one year, as Greenblatt prescribes.

This brings me to 25 stocks in the Magic Formula portfolio, which means I am now fully invested. I started on 7 July 2025 so in my update next month, I will be selling the first 3 names and add 3 new names.

That’s it for this week! Please like or share if you enjoy reading Guru Gems.

You can follow me on X @guru_gems and Substack @gurugems for more insights.

Until next week!

Share Guru Gems

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

Welcome back!

After several longer deep-dives in recent weeks, I wanted to keep this week’s issue short and focused.

Here is what’s on the menu for today:

• 1️⃣ A quick update on the Guru Gems portfolio

• 2️⃣ A look at the current watchlist

• 3️⃣ One interesting thing I read this week

Start here if you are new to Guru Gems.

💎 Guru Gems Portfolio update

The Guru Gems portfolio currently includes 13 stocks and a 26% cash position.

At the end of March, the portfolio was down nearly 6%, after a perfect storm of SaaSpocalypse and Luxury sell-off had hit the portfolio.

Since then there has been a nice rebound as can be seen from the chart below.

The portfolio is up 4% since I started Guru Gems a bit more than a year ago. Not exactly a spectacular result, but also not terrible…

The bounce back was driven by Constellation Software (+15% from end March), Amazon (+30% from end March) and who else than the ultimate Guru Gem Alphabet (+32% since the start of Q2).

Over the past month, there were only 2 ‘activities’ in the portfolio:

• I trimmed Alphabet a little, keeping the position capped to 15% of the portfolio

• The only new position this past month was Zoetis. Read more below on why I added this global leader in animal health:

Subscribe now

👀 Watchlist update

Following the latest 13F releases a few weeks ago, I updated the Guru Gems watchlist, which now includes more than 35 companies that are in Gurus’ portfolios and that may be interesting to study in more depth.

A few names which I’m watching very closely:

• Interactive Brokers (IBKR)

• Netflix (NFLX)

• MercadoLibre (MELI)

• Molina Healthcare (MOH) (part of my Magic Formula portfolio)

Interactive Brokers (IBKR) is a company I reviewed in more detail in last week’s update. At the right price, this is a company I would definitely like to add to the Guru Gems portfolio.

Check out last week’s post if you haven’t yet, where I cover Marc Werres, a lesser-known Guru who has compounded at ~20% per year over the past decade.

📖 One Interesting Thing I Read This Week

I re-read Warren Buffett’s 1999 Berkshire Hathaway shareholder letter, published in March 2000, near the peak of the dot-com bubble.

At the time, Buffett was underperforming badly and many people believed he had “lost it” because he didn’t participate in the internet mania.

History doesn’t repeat itself, but it often rhymes

Back to 2026. Berkshire is again significantly underperforming the S&P 500, by more than 30% over the past year, having ‘missed out’ on the rally driven by mega-cap AI and semiconductor winners.

Has Berkshire lost it?

Or could it be that we are in a phase of the cycle that rhymes a lot with what happened in 1999 and early 2000?

Here are some numbers that should at least make investors think about whether this time is really different:

• Semiconductor index (SOX) up 175% over the past year and 81% year-to-date

• Intel, Micron (chart shown below), AMD, Dell, ASML are within 5-10% of their all-time high

• NASDAQ 100 up by more than 8% in May

• Korea Composite Stock Price Index (KOSPI), in which Samsung and SK Hyninx make up more than 40% of the index, is up 214% over the past year

With the current context in mind, now read the extracts below from Buffett’s 1999 letter (and I highly recommend to read the full letter).

To me, this sure rhymes with the current environment.

(Bold emphasis is mine)

The numbers on the facing page show just how poor our 1999 record was. We had the worst absolute performance of my tenure and, compared to the S&P, the worst relative performance as well.

[…]

We made few portfolio changes in 1999. As I mentioned earlier, several of the companies in which we have large investments had disappointing business results last year. Nevertheless, we believe these companies have important competitive advantages that will endure over time. This attribute, which makes for good long-term investment results, is one Charlie and I occasionally believe we can identify. More often, however, we can’t — not at least with a high degree of conviction. This explains, by the way, why we don’t own stocks of tech companies, even though we share the general view that our society will be transformed by their products and services. Our problem — which we can’t solve by studying up — is that we have no insights into which participants in the tech field possess a truly durable competitive advantage.

[…]

Our reservations about the prices of securities we own apply also to the general level of equity prices. We have never attempted to forecast what the stock market is going to do in the next month or the next year, and we are not trying to do that now. But, as I point out in the enclosed article, equity investors currently seem wildly optimistic in their expectations about future returns.

[…]

Meanwhile, if anyone starts explaining to you what is going on in the truly-manic portions of this “enchanted” market, you might remember still another line of song: “Fools give you reasons, wise men never try.”

As a reminder, Berkshire significantly outperformed the S&P 500 during the market’s ‘lost decade’ of 2000 - 2010…

That’s it for this week! Thank you for reading.

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.

In today’s edition:

• 🕴️ The Guru: Marc Werres — founder of Hinde Group, one of the best investors you’ve probably never heard of

• 🏛 The Playbook: The Hinde Way — The 5 Pillars of a great business

• 📺 The Playbook in action: A Netflix case study (a 4× return in under four years)

• 💎 The Gem: Interactive Brokers Group (IBKR) — Werres’s decade-long compounder

2-minute AI audio version of today’s post:

Guru in the spotlight: Marc Werres

Marc Werres is the founder and Managing Partner of Hinde Group, a San Francisco–based investment firm he launched in August 2015.

His performance over the past decade has been outstanding. Since inception in 2015 through the end of Q1 2026, the portfolio has compounded at ~20% net per year, significantly outperforming the S&P 500’s ~13% return.

He has done this with a concentrated portfolio of less than ten names, no leverage, and through periods where he held meaningful cash when nothing looked attractive.

I came across Hinde Group thanks to a reader recommendation, and the more I read, the more convinced I became that this is a ‘hidden master’ worth highlighting.

Like other great investors, Werres focuses on the long horizon. Hinde holds many positions for years (some since inception) and places almost no weight on short-term benchmarks.

The team’s information edge comes from deep primary research and alternative data, but above all it is the investment discipline characterized by concentration, patience, and strict quality filters that defines the ‘Hinde Way’.

Subscribe now

If you enjoy this newsletter, please help others discover it by clicking 🤍

And if you learned something useful, consider buying me a virtual coffee ☕

The Hinde Way

Hinde Group’s investment approach is built on a small number of clear principles.

1️⃣ Two categories: Compounders & Special Situations

Hinde Group’s investments fall into two categories:

🟡 Compounders are great businesses that are out-of-favor, misunderstood, or underappreciated. The intended holding period is indefinite. Most positions get held for many years.

⚫ Special Situations are less-than-great businesses trading at a meaningful discount to intrinsic value, where a specific catalyst like a spin-off, merger, restructuring, or litigation outcome, is expected to close that gap within two years.

That’s the entire universe. Just compounders to own (most of the portfolio) and special situations to harvest (smaller part of the portfolio).

2️⃣ The 5 pillars of a great business

According to Marc Werres, a great business has all five of the following characteristics

🏰 1. Meaningful & Durable Market Power

Great businesses are those that create large amounts of value for their owners over time. One of the essential elements to achieving that, is meaningful and durable market power.

These businesses are characterized by pricing power, high returns on investment (ROE of 15–20%+), and barriers to entry that protect those returns from competition over the long term.

Pricing power and barriers to entry are also at the core of Chris Hohn’s investment framework:

🔮 2. Predictable Performance

It is hard to overstate how unpredictable the future is. Precisely forecasting the performance of any business over an extended period of time is a fool’s errand.

[…] But it is more reasonable to make such forecasts for some companies than it is for others.

Some companies have certain characteristics that make their performance relatively predictable: a stable market demand, recurring revenue, long product cycles, diverse sources of demand, limited dependencies, flexible operations, high profit margins and conservative financing.

The point is not to forecast quarterly EPS but to be confident that the business will look meaningfully bigger and more profitable in 5–10 years.

Predictability is also the #1 criteria Dev Kantesaria looks for in a business:

🚀 3. Abundant Growth Opportunities

Great businesses have abundant growth opportunities from multiple sources that should allow them to grow earnings per share at a compound annual rate in the high single digits or more over the next decade.

While there are different ways a business can grow (Improving existing products, enhancing pricing effectiveness, reducing costs … ), the most noteworthy way is to develop new products and services, which when done successfully “can create abundant growth opportunities for decades.”

Hinde looks for systematic innovation: maintaining a ‘stock and flow’ of products and services across Core, Emerging, and Viable horizons.

Amazon is a prime example of a company whose ability to successfully develop new products and services seems to provide it a nearly limitless source of growth.

🧠 4. Talented, Honest & Driven Leaders

Leaders who embrace reality, move fast, advocate simplicity, and obsess over attracting top talent. Similar to what I have seen with many other Gurus, they look for owner-operators whose incentives are aligned with long-term shareholders.

🔥 5. A Cult-Like Corporate Culture

Great businesses have cult-like corporate cultures that play an essential role in their success. Those cult-like corporate cultures share three fundamental characteristics:

• Clear & Compelling Vision

• Permeating & Guiding Ideology

• Zealot Employees

You can read more about each of these pillars on Hinde Group’s website.

3️⃣ Concentration with a margin of safety

Hinde Group typically holds about 10 positions or less.

This is a deliberate concentration that focuses capital on their best ideas. At the end of Q1 2026 the portfolio was just seven long positions plus cash.

At the same time, they insists on a margin of safety for every position: a discount relative to a conservative appraisal of intrinsic value. The combination of high concentration and high selectivity is the engine.

We do not need every investment to be successful for the overall portfolio to deliver exceptional results. We only need to limit unsuccessful investments in number and magnitude. Hinde Group’s investment process is designed to target a 90% success rate over time and to ensure unsuccessful investments are not disasters. Even if everything is working exactly as planned, 1 in 10 of our investments will fall short of our 15% hurdle rate when all is said and done.

At the same time, just because we expect to have some unsuccessful investments does not mean that we should just brush them off when they occur. One of Hinde Group’s principles is ‘Own the Outcome’. Putting that principle into practice means recognizing unsuccessful investments when they occur, critically examining what mistakes may have been made, and learning whatever we can from the experience.

🎬 The Hinde Playbook in Action: Netflix (2022–2026)

Before we get to the current portfolio, I want to take a brief detour through what I believe is a brilliant illustration of The Hinde Way.

Werres has a favorite Warren Buffett quote:

“A great investment opportunity occurs when a marvelous business encounters a one-time huge, but solvable problem.”

Netflix in early 2022 was, in Werres’s words, an “archetypal example” of that pattern.

Act 1: Building a position into a 75% drawdown

Hinde began building its NFLX position in Q1 2022, when the stock had fallen sharply from its November 2021 high. They kept adding right through the brutal April 2022 earnings crash that ultimately took the stock to nearly 75% off that all-time high.

The investment community had largely written Netflix off after just a couple of quarters of softer membership growth.

Werres ran Netflix through the 5-pillar framework and concluded it was the rare company that scored well on all of them: meaningful and durable global-scale market power; predictable, low-churn subscription revenue; a vast unpenetrated TAM (only ~6–7% of US TV viewing time, even in its most established market); an elite founder-led management team in Reed Hastings and Ted Sarandos; and a famously cult-like corporate culture.

Some of the challenges that were feared by the market he saw as solvable.

A key signal for Werres was also the fact that Reed Hastings personally bought $20 million of stock in the open market at roughly the same time Hinde was building its position.

Act 2: A strong contrast with Bill Ackman

Werres was not the only great investor who saw Netflix in early 2022 as a “marvelous business with a one-time, solvable problem.”

Bill Ackman saw exactly the same setup.

In January 2022, at almost the same moment Werres was building his Hinde position, Pershing Square put over $1.1 billion into Netflix at similar prices, for almost identical reasons.

Then came the April 2022 earnings report.

Netflix announced it had lost 200,000 subscribers (its first decline in over a decade) and forecast a steeper drop ahead. The stock crashed more than 25% in a single day.

Ackman sold the entire position the very same day, locking in a ~$400 million loss after just three months of ownership.

I covered this full episode in Bill Ackman & The Pershing Square Playbook 👇

As I highlighted in my Ackman post, being “right” about the quality of a business clearly doesn’t help if you sell too soon.

[Quick sidenote: it looks like Bill Ackman and Marc Werres have a very similar investment framework since quite a few names have appeared in both investors’ portfolios at similar times: Netflix, Alphabet, Amazon, Uber, …]

Final act: The payoff

The stock fell further in 2022 before having one of the greatest recoveries of the decade.

By early 2025, NFLX had climbed above $1,000 and became Hinde’s second-largest position at just under 18% of the portfolio.

And this is when Werres started reducing his position:

“Netflix remains a great business with a bright future, but its stock clearly offers a notably lower prospective return above $1,000 than it did below $300.”

Hinde began trimming at the start of 2025, and by the end of Q2 the position was down to just ~2% of the portfolio.

Over a 41-month holding period, the Netflix investment delivered a 62.7% IRR and a 4.1x MOIC (Multiple on Invested Capital, i.e., Hinde got 4.1× their money back).

Bought when the business was great but the market was terrified. Held while the thesis was proven out. And trimmed when the prospective return no longer cleared the hurdle.

Perfect execution…

The sequel: Act 1, Take 2

As I was finishing this piece, Hinde Group released its Q1 2026 letter and Werres has started buying Netflix again.

NFLX is now down more than 30% from its June 2025 high, weighed down by AI-related disruption fears and the company’s failed bid to acquire Warner Bros. Discovery.

The Q1 2026 letter confirms that Werres used the weakness to increase both his Netflix and Amazon positions.

He is not alone. As I covered last week in What Gurus are buying and selling — Q1 2026 13F updates, several other long-term Gurus were also adding to Netflix during the quarter:

• Josh Tarasoff initiated a new ~8% position

• Thomas Russo added +12%

• Bill Nygren significantly increased Oakmark’s stake

The shared thesis is that Netflix is entering a new phase of pricing power and margin expansion, with or without transformative M&A.

In other words, we are watching the Hinde playbook run a second time on the same business with the same framework, but as a new opportunity.

The Hinde Group Portfolio (Updated Q2 2026)

At the end of Q2 2026, the Hinde portfolio consisted of just eight long equity positions and ~9% cash.

“Our portfolio is off to a strong start thus far in the third quarter. I feel good about the individual positions in the portfolio and the portfolio’s overall composition. Nonetheless, I always aim to identify great new investments that will improve the portfolio even further. I remain optimistic about how our portfolio will do in the coming years.”

There are no exact weightings available of the positions, but from the letters we can get a pretty good idea of what the portfolio looks like:

Interactive Brokers Group (IBKR), Uber Technologies (UBER), Northeast Bank (NBN), Alphabet (GOOG)

• At the start of 2025 these 4 companies accounted for most of the portfolio’s equities

• Werres mentioned some trimming and adding during 2025, but it looks like these are still significant positions in the current portfolio

• In the Q1 2026 letter, Werres mentions that IBKR “still represents a mid-teens percentage of the portfolio’s value”

Amazon.com (AMZN), Netflix (NFLX)

• Positions that had been trimmed during 2025

• Werres has been adding again to both names in Q1 2026

• Amazon began 2026 as one of their smallest positions, then significantly increased their position in Feb at an avg. price of ~$203 after the shares sold off

“AMZN ended the second quarter at $237.35. Although the stock had a nice pop in late July after reporting results for the second quarter and currently sits at $274.48, it remains meaningfully undervalued. The shares still offer a prospective annualized return north of 20.0% over a long time horizon.”

Becton, Dickinson and Company (BDX)

• BDX is a ‘special situation’ stock added in Q3 2025

• It was their first new position since 2022.

New position (Not yet disclosed)

“Essentially all of the new position that Hinde Group began to initiate at the very end of the second quarter was purchased at the beginning of the third quarter. That position will be covered in the partner letter for the third quarter.”

Three observations from this portfolio:

1/ Having the patience to wait for 2 years before buying something new, reminds me of Norbert Lou’s Punch Card mindset:

2/ Several names overlap with the Guru Gems portfolio (GOOG, UBER, AMZN), which is a great validation :)

3/ Finally, the portfolio’s evolution over the past decade is also quite fascinating.

• Werres bought IBKR at inception in 2015.

• He added Alphabet in 2017 and UBER in 2020.

• He initiated Netflix during the 2022 valuation collapse and harvested most of it in 2025.

• Werres is comfortable holding meaningful cash when attractive opportunities are scarce. In Q1 2017, after some big wins, cash was nearly at 48% of the portfolio.

In summary: The discipline is to ride winners while constantly hunting for the next mispriced great business or special situation and having the patience to wait if nothing good shows up.

💎 The Gem: Interactive Brokers Group (IBKR)

Werres has owned IBKR since Hinde Group’s first day. The position has delivered a ~27% compound annual return over more than a decade.

IBKR is the position he keeps returning to in his letters, and IBKR was the largest single contributor to Hinde’s 2025 performance.

So if there is one Gem worth studying from this portfolio, this is this one.

A brief callback: Bryan Lawrence

Long-time readers may remember that I touched on IBKR last summer in The Lawrence Method, where Bryan Lawrence’s Oakcliff Capital held it as their largest position at over 22% of portfolio (now 30%).

Lawrence’s four-question framework gave IBKR a green check on three of four boxes (Business We Understand, Good Business with Durable Cash Flows, Shareholder-Aligned Management) and a question mark on valuation.

When I wrote about Lawrence in July 2025, I concluded that while IBKR looks like a high quality business, I should do a bit more research on this company, and that’s exactly what I did when researching for today’s Guru Gem.

What does Interactive Brokers actually do?

IBKR is a highly automated global electronic brokerage firm founded by Thomas Peterffy. It provides ultra-low-cost trading across stocks, options, futures, FX and other assets in 150+ markets.

Key customers are primarily professional traders, hedge funds, and financial advisors, but also include active retail traders.

The business has two main revenue streams:

• Commissions earned on trades. In 2025, commission revenue grew 26%.

• Net interest income earned on customer credit balances and margin loans. In 2025, this grew 13% despite multiple Fed rate cuts, because customer credit balances grew 28% and margin loans grew 31%.

IBKR’s Q1 2026 results show how impressive this business really is:

• $1.7 billion in net revenue (+17% YoY)

• $1.3 billion in pre-tax income (+22% YoY)

• 77% pre-tax margin (highest profit margin in the industry)

• 4.8 million customer accounts (+31%)

• $789 billion in client equity (+38%)

Why IBKR is a ‘great business’ by Hinde’s 5-pillar framework

🏰 1. Meaningful & Durable Market Power ✅

IBKR’s automation-first cost structure lets it offer what Werres calls “the least expensive brokerage services while maintaining the highest profit margins in the industry.”

In his Q3 2019 letter he highlighted that IBKR was charging 3.40% on a $25,000 margin loan while Schwab and Fidelity each charged 8.82% for the same loan. Today, per IBKR’s Q1 2026 investor materials, US margin rates are 4-5% for IBKR Pro, still well below larger retail brokers.

When Schwab and TD Ameritrade went to zero-commission in late 2019, Werres argued in that same letter that while this would hit some online brokers hard, IBKR would be “relatively unscathed”, which turned out to be a correct analysis.

🔮 2. Predictable Performance ✅

IBKR’s ‘Total Customer Equity’ has grown at a 25% annualized rate over the last decade, proving its intrinsic value grows regardless of the macro environment.

This is the metric Werres started emphasizing in his 2023 letters because it captures intrinsic value growth without the noise of interest rates and trading volumes.

Customer accounts have grown alongside it at a similar high pace.

🚀 3. Abundant Growth Opportunities ✅

IBKR has penetrated less than 1% of its addressable market.

Growth levers include introducing-broker partnerships (the “white-label” model that fueled Asian growth), new geographies and new products (crypto, prediction markets via ForecastEx, overnight trading, mutual funds, bonds, event contracts).

🧠 4. Talented, Honest & Driven Leaders ✅

Founder Thomas Peterffy is still chairman, still owns roughly 73% of the underlying LLC, and still drives the company. Senior management is dominated by software engineers, not bankers. And the company emphasizes long-term growth over short-term metrics.

🔥 5. A Cult-Like Corporate Culture: ✅

The mission has not changed since 1977:

“Create technology to provide liquidity on better terms. Compete on price, speed, size, diversity of global products and advanced trading tools.”

Every product decision like the recent push into prediction markets and AI tools, runs through that same lens. This is the kind of cultural continuity that tends to survive the founder.

Five out of five… looks like we found ourselves a ‘good’ business :)

To Buy or Not to Buy

Now that we’ve established that IBKR is a great business, let’s look at the valuation

IBKR was up 113% in 2024, 46% in 2025 and currently trades at roughly 36x earnings.

Has it become too expensive?

The fact that IBKR is still a significant position in the portfolios of Bryan Lawrence and Marc Werres, indicates that they still expect to earn a decent return from this stock, even at the current levels.

Here is what Werres wrote in his Q4 2025 letter:

“Interactive Brokers should still grow its earnings at least at a mid-teens annual rate. […] Interactive Brokers also aims to pay out a dividend equivalent to a 0.5% to 1.0% yield. Between earnings growth and its dividend yield, IBKR seems poised to deliver a high teens or better annualized return over the coming years. IBKR would need to trade at more than $100 per share today to offer a market-level return.”

At the price IBKR traded at end of 2025 ($65-70), Werres is implicitly saying IBKR still offers above-market prospective returns, even after the recent run.

At the same time, Werres indicated in his latest letter that he sold some IBKR (probably in the $75-80 range), not because the thesis broke, but because the price had run hard:

While IBKR gained only 4.3% for the quarter as a whole, we trimmed some of our position early in the quarter at prices notably higher than where IBKR ended the quarter

Final decision

📌 Keeping IBKR on the watchlist (but moving it to the top of the list)

The business is exceptional, the long-term thesis is intact, but entry price matters as we have shown with the Netflix case study.

The long-term thesis is supported by two of the most disciplined long-term investors I have studied, who have both built their portfolios around this stock:

• Bryan Lawrence (Oakcliff Capital): IBKR held since 2017, now 30% (!) of portfolio

• Marc Werres (Hinde Group): IBKR held since 2015, ~15% position, ~27% CAGR over the holding period

That is a meaningful long-term endorsement.

My entry trigger: I will initiate a position if IBKR pulls back to around $65 or below, which is the Q1 2026 closing level and would put it materially below Werres’s trim range and meaningfully below his $100 market-level-return threshold.

That’s it for this week’s edition!

Thank you for reading and following along.

Until next week!

Disclaimer: This newsletter is for informational purposes only and does not constitute financial advice or investment advice. Always do your own research before making any investment decision.