Ask Michael Mauboussin and Dan Callahan whether negative FCF is good or bad, and they answer in two words: “It depends.”²

Keep Walmart in mind. In its last fiscal year it took in $713.2 billion, every grocery cart and flat-screen TV bought by 280 million weekly shoppers across more than 10,900 stores.³ This year, four companies (Alphabet , Amazon , Meta and Microsoft ) plan to spend close to $800 billion on chips, servers and data centers.⁴ That’s more than all of Walmart’s sales. It works out to about $2.2 billion a day, Sundays included.

The cash is already thinning. By Mauboussin and Callahan’s count, those four plus Oracle produced about $35 billion of FCF over the four quarters to mid-2026. Two years earlier, the figure was $170 billion.⁵

Two managers I’ve followed for years read those numbers through the same equation and landed on opposite sides. David Rolfe and his team at Wedgewood Partners bought more Microsoft, Amazon and Meta in the second quarter. Are the hyperscalers still magnificent? “We certainly think so and have increased our weightings accordingly.”⁶ Fred Liu at Hayden Capital used the same math to argue that the returns everyone is projecting may not hold.⁷

I don’t know who’s right. Neither do they, yet. What I can share is how I’m working through it: five questions that separate their views, the numbers behind each answer, and the signals that should settle the debate over the next 12 to 24 months.

One equation, two readings

Liu writes it in a single line: earnings growth equals return on invested capital (ROIC) times the reinvestment rate.⁸ Earn 20% on your capital, plow back half your profits, and you grow 10% a year. Plow back everything and you grow 20%, provided the 20% holds.

That condition is the whole fight. Growth creates value only while the return on the new dollar beats the cost of capital. Below that line, growth destroys value, however good the revenue chart looks.⁹

Both managers agree the hyperscalers are pushing reinvestment close to 100%. They part ways on what all that new capital will earn, and for how long.

1. Is the return showing up in cash?

This is Wedgewood’s home turf. Its team skips accounting earnings and measures something harder to dress up: new cash flow against new gross assets.

Over two years, Microsoft added an average of $100 billion in assets and $27 billion in incremental cash flow, a return near 30%.¹⁰ A quarter later the numbers were $115 billion and $35 billion. Still almost 30%.¹¹ Amazon is spending about $200 billion this year, enough to enlarge its asset base by an estimated 28%, while its cash flow grows around 30%. Alphabet grew a gross asset base of nearly $450 billion by 40% and kept returns above 30%.¹²

Fred Liu looks at the same quarter and sees the bill. Alphabet spent $44.9 billion on property and equipment in the second quarter, against $39.1 billion of cash from operations. FCF: minus $5.9 billion.¹³ By Liu’s count, it was Alphabet’s first negative quarter since its 2004 IPO.¹⁴

Both readings fit the facts. The tiebreaker is Wedgewood’s own yardstick, tracked quarter after quarter: how many dollars of new cash flow each dollar of new assets produces. Mauboussin’s version is ROIIC, the return on incremental invested capital. If after-tax operating profit rises from $100 to $140 on $200 of new investment, ROIIC is 20%.¹⁵

On consensus forecasts, the four plus Oracle hit a combined ROIIC near 40% in early 2026, against a cost of capital of about 8%. Analysts see it bottoming around 23% in the third quarter of 2027 and recovering to near 35% by 2030.¹⁶ Even the trough clears the bar. What I want to know is whether the trough lands where consensus puts it.

2. Durable return, or scarcity price?

Now Fred is on his home turf. Sell-side models that show 25% to 50% returns on AI projects run off today’s spot price for compute, and new supply could break that price. Tencent said it could resell compute it had ordered months earlier at a markup of more than 30%, which looks more like hoarding than end demand. As SK Hynix, Samsung and CXMT bring capacity online, those premiums should shrink.¹⁷

His line: “The sustainability of ROIC matters more to a stock price, than its absolute level.”¹⁸

The base rates side with the worry. Mauboussin and Callahan found that the gap between the top and bottom fifths of companies by ROIC narrows from 35 points to 15 within five years. In technology, ROIC persistence over five years is weak, a correlation of 0.25. And 15% of top-quintile companies drop to the bottom quintile within three years, where shareholders lose about 11% a year.¹⁹

Wedgewood’s reply is cost. Alphabet says its in-house TPU chips cut the cost of serving Gemini by almost 80%.²⁰ A company that builds capacity cheaper than anyone can rent it will be the last to feel falling compute prices.

That’s why I’d track ROIIC company by company. Consensus already expects Meta’s to fall from about 70% to 16% by late 2027.²¹ An average can hide a weak member for years.

3. Who’s paying?

A return earned on internal cash and the same return earned on borrowed money carry different risks.

Fred Liu tracks the shift. Debt went from funding about 9% of hyperscaler capex in 2024 to roughly 32% over the last twelve months.²² Equity joined in June, when Alphabet launched an $80 billion stock sale, upsized to $84.75 billion at pricing, with $10 billion of it sold to Berkshire Hathaway.²³ Meta’s Q3 report, out on October 21, shows the same pattern: negative FCF of $2.12 billion, no buybacks so far in 2026, and long-term debt up from $58.7 billion to $80.9 billion since December.²⁴

Daniel Gladiš at Vltava describes the broader move from buying back stock to issuing it.²⁵ GMO’s Ben Inker and John Pease put a price on it. For the first time in twenty years outside the financial crisis, issuance exceeds buybacks, and historically every 1% increase in equity supply has cut total market value by about 4% over the following year.²⁶

Wedgewood points to an offset most investors skip: the hyperscalers’ stakes in AI labs. Alphabet owns about 14% of Anthropic. Microsoft’s OpenAI stake and Amazon’s Anthropic stake are each worth $100 billion or more, by Wedgewood’s estimates. Meta’s AMD warrants are worth close to $90 billion, about nine times its extra memory bill for 2026. Alphabet’s SpaceX and Anthropic gains alone, about $150 billion if realized, could pay for several years of extra memory-driven capex.²⁷ None of it shows up in reported earnings.

The question I’d ask about that hedge is simple: what is it worth on the day you need it? Those stakes carry private marks set by the same AI cycle. If the cycle turns, they shrink just as the capex bill comes due. Whether any of them turns into cash before then is one of the cleaner signals in this debate.

4. Are today’s earnings normal?

This one reaches past the hyperscalers to the earnings of the whole market.

Gladiš describes a timing gap. When a hyperscaler buys a server, the supplier books the sale today. The buyer spreads the cost as depreciation over the next several years. The cycle lifts profits now and mails the bill later. S&P 500 net margins sit near 15%, against a historical norm in the single digits. Record profits aren’t the issue, he writes: “The problem is mistaking them for the norm.”²⁸

Jacobo Arteaga and Javier López Bernardo at BrightGate Capital reach the same place through Michał Kalecki’s profit equation, in which one company’s investment is another company’s profit. During the 1998–2001 telecom and internet buildout, S&P 500 operating earnings climbed 43% from late 1998 to mid-2000, then fell almost 40% in under a year. “It is telling that the bubble burst precisely at the peak of earnings.”²⁹

Bill Nygren at Oakmark asks the question at the supplier level. Micron’s projected 2027 operating margin is above 80%, while its history runs in the high single digits. Oakmark stayed away from AI hardware because it doubted those earnings would last, and it paid for that. AI-related names produced more than half of the Russell value index’s performance this year, and Nygren admits Oakmark missed them.³⁰

Nvidia shows how slippery “normal” gets. Bill puts it at about 11x consensus 2028 earnings.³¹ Liu puts it at about 21x current earnings, below Costco’s 43.³² Neither calls it expensive. Read together, they’re saying the multiple is only cheap if the earnings are real.

Even Wedgewood, the bull in this story, keeps its distance from chipmakers. It warns that semiconductors are cyclical and that peak earnings deserve trough valuations.³³ Its optimism is about the companies buying AI capacity. The sellers are another matter.

5. What if consensus is wrong?

Every reassuring number in question 1 is an analyst forecast. Mauboussin flags it himself: estimates grow less reliable the further out they reach.³⁴

The most revealing table in his paper tracks revisions from August 2025 to August 2026. Analysts raised their 2027 capex forecasts by $210 billion at Alphabet, $142 billion at Amazon, $116 billion at Microsoft and $86 billion at Meta. They raised 2027 operating income by $52 billion, $22 billion, $15 billion and $2 billion. Nvidia and Micron, the suppliers, got $231 billion and $192 billion of upgrades.³⁵

So far, the buyers absorb the capex and the sellers book the profit.

Small shifts in that picture count for more than they seem. In Mauboussin’s math, if expected growth slips from 10% to 7%, next year’s earnings drop only 2.7%, but the justified P/E falls 22.9%.³⁶ Rates pile on. On October 5 the 10-year Treasury yield closed at 5.31%,³⁷ and the Nasdaq 100 hit a record the same day. Long-duration assets should feel a higher cost of capital, Wedgewood notes, “yet very few AI stocks are punished as such.”³⁸

Oakmark, the fund that missed the run, also spells out the record any skeptic has to beat. For the largest tech companies, “earnings growth not only failed to regress—it accelerated.”³⁹ Their 2026 consensus earnings are about twice what analysts expected back in 2019.

What to watch for through 2027

• New cash flow per dollar of new assets: Wedgewood’s own measure. If Microsoft and Amazon keep converting near 30%, the bull case holds.

• The ROIIC trough: consensus expects about 23% in the Q3 of 2027, against an 8% cost of capital.⁴¹ A deeper or later trough favors Fred Liu.

• Operating income revisions vs. capex revisions: if earnings forecasts don’t catch up, the buyers are funding the suppliers’ record year.

• Compute and token prices: falling prices with steady volumes would test Liu’s scarcity argument head-on.

• Debt share of capex funding and net share issuance: both rising favors the skeptics.

• Depreciation against revenue at the hyperscalers: the deferred bill from question 4 shows up here first.

• June 12, 2027: the last SpaceX lockup expires, freeing roughly $2 trillion of stock, with Anthropic and OpenAI listings possibly landing in the same window.⁴²

• Meta: consensus already expects its incremental returns to fall the furthest, and its Q3 brought negative FCF and more debt. If the bear case is right, it should show there first.

Picture an investor in 1980 holding Walmart’s cash flow statement. Seven years of red ink behind it, six more still to come. What mattered was never the cash flow. It was what each dollar would earn later. Wedgewood and Hayden are asking that question about $800 billion. By 2028 we should know which kind of red ink this was.

DISCLAIMER: This post summarizes the views of the investment managers and analysts cited. Their opinions are their own, not those of Ezra's Notes or its author. Figures come from their published letters and research and from company filings, as listed in the notes; consensus estimates can change. I have no position in any of the companies mentioned. This post is for educational and informational purposes only. It is not investment advice or a recommendation to buy or sell any security. Do your own research or consult an advisor before making any investment decision.

Notes

1. Michael J. Mauboussin and Dan Callahan, “To Free or Not to Free (Cash Flow),” Consilient Observer, Counterpoint Global, Morgan Stanley Investment Management, September 17, 2026. Consensus figures are FactSet estimates as of August 31, 2026, and cover Amazon, Alphabet, Microsoft, Meta and Oracle. https://www.morganstanley.com/content/dam/im/assets/publication/thought-leadership/consilient-observer/article_tofreeornottofree_ltr.pdf

2. Ibid.

3. Walmart Inc., Form 10-K for the fiscal year ended January 31, 2026 (total revenues of $713.2 billion; approximately 280 million customers a week in more than 10,900 stores). https://www.sec.gov/Archives/edgar/data/104169/000010416926000055/wmt-20260131.htm

4. Hayden Capital, Quarterly Letter Q2 2026, August 27, 2026, Fred Liu. https://content.haydencapital.com/Hayden-Capital-Quarterly-Letter-2026-Q2.pdf

5. Mauboussin and Callahan, “To Free or Not to Free (Cash Flow).”

6. Wedgewood Partners, Second Quarter 2026 Client Letter, “Are Hyperscalers Still Magnificent?”, July 2026, David Rolfe, Michael Quigley and Christopher Jersan. https://wedgewoodpartners.com/wp-content/uploads/2026/07/Wedgewood_2Q2026_Hyperscalers.pdf

7. Hayden Capital, Quarterly Letter Q2 2026.

8. Ibid.

9. Michael J. Mauboussin and Dan Callahan, “The Math of Value and Growth: Growth, Return on Capital, and the Discount Rate,” Consilient Observer, June 9, 2020. https://www.morganstanley.com/im/publication/insights/articles/article_themathofvalueandgrowth.pdf

10. Wedgewood Partners, Second Quarter 2026 Client Letter.

11. Wedgewood Partners, Third Quarter 2026 Client Letter, “Bad Moon Rising,” October 2026, David Rolfe, Michael Quigley and Christopher Jersan. https://wedgewoodpartners.com/wp-content/uploads/2026/10/Wedgewood-Partners-Third-Quarter-2026-Client-Letter-Bad-Moon-Rising-2.pdf

12. Wedgewood Partners, Second Quarter 2026 Client Letter.

13. Alphabet Inc., Form 8-K, Exhibit 99.1, Q2 2026 earnings release, July 2026 (net cash provided by operating activities $39,069 million; purchases of property and equipment $44,924 million; free cash flow $(5,855) million). https://www.sec.gov/Archives/edgar/data/0001652044/000165204426000066/googexhibit991q22026.htm

14. Hayden Capital, Quarterly Letter Q2 2026.

15. Mauboussin and Callahan, “To Free or Not to Free (Cash Flow).”

16. Ibid.

17. Hayden Capital, Quarterly Letter Q2 2026.

18. Ibid.

19. Michael J. Mauboussin and Dan Callahan, “ROIC and the Investment Process: ROICs, How They Change, and Shareholder Returns,” Consilient Observer, June 6, 2023. https://www.morganstanley.com/im/publication/insights/articles/article_roicandtheinvestmentprocess.pdf

20. Wedgewood Partners, Second Quarter 2026 Client Letter.

21. Mauboussin and Callahan, “To Free or Not to Free (Cash Flow).”

22. Hayden Capital, Quarterly Letter Q2 2026.

23. Alphabet Inc., Form 8-K, Exhibit 99.1, “Alphabet Announces Proposed $80 Billion Equity Capital Raise,” June 1, 2026. https://www.sec.gov/Archives/edgar/data/0001652044/000119312526257724/d83560dex991.htm. Upsizing to $84.75 billion at pricing: Cleary Gottlieb, “Alphabet in $80 Billion Equity Offerings,” June 4, 2026. https://www.clearygottlieb.com/news-and-insights/news-listing/alphabet-in-80-billion-equity-offerings-jun-2026

24. Meta Platforms, Inc., Form 8-K, Exhibit 99.1, Q3 2026 earnings release, October 21, 2026 (free cash flow of negative $2.12 billion; long-term debt of $80.92 billion at September 30, 2026, against $58.74 billion at December 31, 2025). https://www.sec.gov/Archives/edgar/data/0001326801/000162828026066700/meta-09302026xexhibit991.htm

25. Vltava Fund SICAV, Letter to Shareholders 3/2026, “Tadej Pogačar and investing in stocks,” October 2026, Daniel Gladiš. https://www.vltavafund.com/dopisy-akcionarum/tadej

26. Ben Inker and John Pease, “A Catalyst for the AI Bubble Break,” GMO Quarterly Letter, 3Q 2026. https://www.gmo.com/americas/research-library/a-catalyst-for-the-ai-bubble-break_gmoquarterlyletter/

27. Wedgewood Partners, Second Quarter 2026 Client Letter.

28. Vltava Fund, Letter to Shareholders 3/2026.

29. BrightGate Capital SGIIC, Interim Letter to BrightGate Global Income co-investors, first half of 2026, Madrid, July 13, 2026, Jacobo Arteaga Fierro and Javier López Bernardo. https://www.brightgatecapital.com/documentacion/brightgate/global-income/cartas/2026-H1-ENG.pdf

30. Harris | Oakmark, “A Conversation with Shareholders,” U.S. Equity Market Commentary 3Q26, September 30, 2026, Bill Nygren. https://oakmark.com/wp-content/uploads/sites/3/2026/10/A-conversation-with-shareholders_U.S.-equity-market-commentary_3Q26_vF.pdf

31. Ibid.

32. Hayden Capital, Quarterly Letter Q2 2026.

33. Wedgewood Partners, Second Quarter 2026 Client Letter.

34. Mauboussin and Callahan, “To Free or Not to Free (Cash Flow).”

35. Ibid.

36. Mauboussin and Callahan, “The Math of Value and Growth.”

37. U.S. Department of the Treasury, Daily Treasury Par Yield Curve Rates, October 2026. https://home.treasury.gov/resource-center/data-chart-center/interest-rates/TextView?type=daily_treasury_yield_curve&field_tdr_date_value_month=202610

38. Wedgewood Partners, Third Quarter 2026 Client Letter.

39. Harris | Oakmark, “A Conversation with Shareholders,” 3Q26.

40. Earnings dates from the companies’ announcements. Alphabet: “Alphabet Announces Date of Third Quarter 2026 Financial Results Conference Call” (October 28, 4:30 p.m. ET). https://abc.xyz/investor/news/news-details/2026/Alphabet-Announces-Date-of-Third-Quarter-2026-Financial-Results-Conference-Call-2026-8tpGZsLS6v/default.aspx. Microsoft: “Microsoft announces quarterly earnings release date” (October 28, after market close). https://finviz.com/news/400058/microsoft-announces-quarterly-earnings-release-date. Amazon: “Amazon will share its Q3 2026 earnings on October 29.” https://www.aboutamazon.com/md/news/company-news/amazon-earnings-q3-2026-report

41. Mauboussin and Callahan, “To Free or Not to Free (Cash Flow).”

42. Inker and Pease, “A Catalyst for the AI Bubble Break.”

    ---

    ---

    Which reading do you find more convincing, Wedgewood's or Hayden's? And which signal on that list would change your mind? Tell me in the comments.

    And if you liked this post, share the love!

Share

So he started buying companies, the kind other owners wanted to get rid of. He began with €50,000 and acquired ten businesses in three years. Today he invests through Bavaria Industries AG, and what stands out in his writing is what he chooses to dwell on. Not the wins. The errors.

His biggest one fits in a single sentence. Russian shares that looked cheap by every conventional measure, about 5% of his portfolio, became unsellable almost overnight. He treated the war as too irrational to happen, and the loss changed how he sizes every position since.

In this interview, Reimar talks about how a failed job search became a company group, why he starts every idea by asking how it could go wrong, a UK franchisor he keeps adding to, and the seven questions he runs before buying anything.

One of them comes before all the others: “Can we survive being completely wrong?”

Tell me about your background and how you got into investing.

At 36, I hit rock bottom. After numerous job changes, a headhunter offered me a position only as a recruitment consultant, an unmistakable sign that I was no longer placeable as an employee. Only one path remained: to become an entrepreneur myself.

Without my own business idea, the only option was to buy an existing company. For two years I searched from home, relying on vague financing promises, for a suitable takeover candidate. The turning point came after a short internship with a professional company acquirer. There I learned to hunt specifically for businesses that others wanted to get rid of: true problem cases.

We started with €50,000 in equity and acquired ten companies in the first three years alone. Since then, through numerous acquisitions and investments, we have built something substantial. That is primarily the result of countless errors. Every false assumption and every failed investment forced us to re-examine our processes, and from that a system emerged, the BAVARIA Operating System, which has long-term outperformed the DAX and the S&P 500.

What’s your investment approach, and how has it evolved over time?

My focus has moved from upside to error reduction. Most ambitious investors, and I was one of them, spend their time looking for the next Microsoft or Nvidia and too little on what could go wrong. So we reversed the normal question. Besides “Why will this company succeed?”, we ask “What could prevent it from succeeding?”

Every mistake has to leave something behind: a rule, a warning sign or a better question for the checklist. A carmaker improves quality by tracking defects, and investing should work the same way. If you systematically reduce the number of losers, the winners naturally take up more room.

Some practices follow from that. For new or riskier investments we start small and average in, because a small first position gives us time to learn and stops us from defending a decision instead of analysing it. We write our theses down, keep an investment journal as a defence against self-deception, and sleep on decisions. And we prefer solid growth at a sensible price to expensive dreams. We hold very few classic growth stocks, apart from an Alphabet position we exited too early.

The Russian loss also changed how we think about country risk. In emerging markets we now spread our bets more widely and take some profits along the way. Over the past 20 years our returns have exceeded the S&P 500. That doesn’t prove we are unusually clever, but it suggests that avoiding serious mistakes can take you surprisingly far.

Tell me about one of your biggest investment mistakes, what happened, and what did you take away from it?

Our biggest mistake was Russia. Almost overnight, our Russian shares became effectively worthless. That was around 5% of our portfolio. Keep it in perspective: we lost capital, while people in Ukraine lost their lives.

We made two errors. First, we massively underestimated the probability of a Russian attack. We didn’t think war was impossible, but because the outcome looked irrational for everyone involved, we treated it as remote. Investors like to assume that leaders act in their economic self-interest, and politics doesn’t work that way. Second, we underestimated how severe and how fast the financial sanctions would be. Shares that looked extremely cheap became unsellable within days, so the value on the spreadsheet no longer mattered.

Shortly before the war, I was arguing at an investor conference that emerging markets were attractive, with about 25% of our portfolio invested there. My conclusion was almost embarrassingly simple: “What can really go wrong at these valuations?” We now know the answer: a lot.

There was one more lesson. Our original investment was only about 2% of the portfolio, and the rest was accumulated gains. But profits are capital once they exist, so position size should be measured against current portfolio value, not historical cost. We should have trimmed the position after the strong gains.

Since then, we ask different questions before investing.

• Can foreign investors enforce ownership rights?

• Can capital leave the country?

• Could sanctions or war make the shares unsellable?

• How large should the position be if the worst case is zero?

• The first question now is: can we survive being completely wrong?

What’s one idea on your radar right now? Walk me through it

One we recently added to is The Property Franchise Group (TPFG), a UK franchisor of estate agency and lettings brands. We hold about 3% of our portfolio in it and own 5% of the company.

TPFG doesn’t own branches. Local offices belong to franchisees, and TPFG provides brand, technology, training and support in exchange for recurring royalties. It runs the largest franchised property network in the UK, with more than 930 branches and roughly 153,000 managed rental properties, yet it still holds only about 3% of a very fragmented market. Roughly 70% of its earnings come from franchising, and around 70% of those fees are linked to lettings rather than house sales. Lettings are far more stable, which is why earnings kept growing through 2008–2009 and the 2022–2023 slowdown.

The economics are strong: EBITDA margins of about 45%, free cash flow conversion above 100%, and after-tax returns on invested capital above 20%. Since its 2013 listing, adjusted earnings per share have compounded at around 23% a year. Management has also been disciplined. It kept acquired brands intact after the 2024 Belvoir merger and focused on supporting franchisees.

The stock trades at about 10x expected free cash flow, as the price has dropped further, with net debt of only 0.3x EBITDA. We think EPS can compound at 13–15% a year. The risks are a prolonged weak UK housing market, regulation and acquisition integration. The opportunity exists, I think, because it sits in an unfashionable corner: small-cap UK equities, limited analyst coverage and a sector people associate with cyclicality.

How do you think about concentration versus diversification? And how do you decide when to sell?

Since we invest in developing countries and very exotic markets we cannot concentrate enough as hard as we try. And since our philosophy is to average in and build positions slowly we always have a portfolio of ideas we are testing.

Our biggest 5 positions are more conventional bets and make out roughly a third of our equity portfolio and a quarter of our NAV - as we always hold a quarter or so of cash and gold as counterbalance to our more risky equity bets.

Name a couple of books or podcasts you’ve come across recently that have made you a better investor.

The first is Superforecasting by Philip Tetlock and Dan Gardner, built on the Good Judgment Project, which showed that ordinary people can outperform intelligence analysts.

Superforecasters don’t treat predictions as yes-or-no events but as probabilities. Instead of saying “this company will turn around,” they say 65% and keep updating that number as new information arrives. They also break a big question into smaller parts, separating what they can know about the specific case (the inside view) from general base rates (the outside view).

What I value most is their humility. They keep scorecards of past predictions and treat mistakes as data to calibrate their thinking, not as character flaws. And they combine many perspectives instead of relying on one grand narrative.

The second is Andy Grove’s work, mainly High Output Management and Only the Paranoid Survive. Grove treated management as an engineering discipline. A manager’s output is the output of the units under their influence, so the job is finding the activities with the highest leverage per unit of time.

His idea of task-relevant maturity says there is no single best way to manage: the style has to fit the person’s maturity for that specific task.

He also focused on strategic inflection points, when the old way of doing business dies and survival depends on recognizing the shift early, and he built a culture of constructive confrontation, where data beats hierarchy.

Share your core criteria checklist: the handful of non-negotiables every investment idea has to pass before you’ll consider it.

• Is the balance sheet too weak?

• Do we underestimate cyclicality?

• Do we trust adjusted earnings too much?

• Do we ignore dilution?

• Do we overpay for growth?

• Do we buy without a clear thesis?

• Do we mistake a temporary recovery for a durable improvement?

If you could leave one piece of advice for other investors reading this — something you wish someone had told you earlier — what would it be?

Buying a stock is like buying a used car, always look for a snag and who is the fool in the transaction. Only buy if you identified a fool, if not put it into the too difficult basket (like most ideas).

Idea Snapshot

Reimar’s Core Criteria Checklist

DISCLAIMER: This interview reflects the views of Reimar Scholz alone. They are not the views of Ezra’s Notes or its author. Figures and statements are as provided by the interviewee and have not been independently verified. This post is for informational purposes only and is not investment advice or a recommendation to buy or sell any security. Do your own research before making any investment decision.

If this conversation made you think, tell me what stayed with you in the comments. And if you know someone who would value it, share it with them.

Share

Steve Jobs once phoned Warren Buffett with one question: what should Apple do with all its cash? Buffett answered. Jobs did nothing.

I remembered that story a few days ago. I had always assumed, and I suspect many readers have too, that Jobs lacked the skills of a good capital allocator. So I went through Apple’s capital allocation during his second run at the company, from 1997 to 2011. I came away thinking that Steve, in his own way, was a good one.

Buffett told the story on CNBC’s Squawk Box in 2012.² Jobs had called him a couple of years earlier.

Warren laid out four choices: pay dividends, buy back shares, make acquisitions, or sit on the cash. There was a lot to sit on. In the months when the call could have happened, Apple held between $39.8 billion and $51 billion.³

To size that pile: with the $51 billion it had in September 2010, Apple could have paid cash for all of BNSF, debt included. That deal came to $44 billion, the largest acquisition in Berkshire’s history, and Warren had closed it that February.⁴ The $51 billion was also more than Germany spent on its military in 2010, about $45 billion.⁵

Buffett told him that if he thought his own stock was undervalued, he would use the cash to buy it back. Then he asked what Jobs thought of Apple’s. Steve said it was very undervalued. They talked a while longer, and Jobs did nothing. “He just liked having the cash,” Buffett said. Jobs kept his mountain of cash, much as Warren keeps one at Berkshire Hathaway.

Buffett later learned that Jobs had been telling people Warren agreed with him about doing nothing.

By Warren’s account, Apple traded at two hundred and something dollars at the time. Adjusted for the two stock splits since, that is $7 to $11 a share in today’s prices.⁶

Buffett gave no date. Counting back a couple of years from 2012 and matching that price, the call falls between October 21, 2009, when the stock first closed above $200, and October 13, 2010, when it crossed $300.⁷

Jobs had reasons to think the stock was cheap. The call came in the iPad’s year: unveiled in January 2010, on sale in April. The iPhone 4 followed in June. iPhone sales had nearly doubled in 2009 and nearly doubled again in 2010.⁸

I think Steve saw three things the market had not priced in:

1. He knew what was coming. He ran the products, so he saw them before anyone else. Apple’s earnings grew 70% in 2010 and 85% in 2011.⁹

2. Apple’s accounting hid part of its profit. Until January 2010 the company did not book an iPhone sale all at once. It spread the revenue over the device’s estimated life. Under that rule Apple earned $5.7 billion in 2009. Without it, $8.2 billion.¹⁰

3. His own role was one of Apple’s big risks. In January 2009 Jobs took medical leave through June. Six days later, in the middle of the financial crisis, the stock hit its low for the year: $78.20. Some of that discount may still have been in the price in 2010.¹¹

At the end of fiscal 2010, net of cash, Apple traded at 15x that year’s earnings and 8x the next year’s.¹²

Steve Jobs as a capital allocator

I wanted to look at Jobs’s second run through capital allocation alone and see whether he was any good at it. When he took over as interim CEO in September 1997, Apple was closing a year with a $1.05 billion loss. It had $1.5 billion in cash and $951 million in long-term debt. The rating agencies had cut that debt to junk, and Microsoft had put in $150 million for preferred stock.¹³

Steve later said Apple had been 90 days from bankruptcy.¹⁴ That is where his capital allocation starts.

What Jobs did

He cleaned up the balance sheet first. In 1998 he paid off what was left of the bank loans. In 1999, $661 million of convertible notes turned into stock. In 2004 he repaid the last $300 million in cash. He issued no new debt in 14 years.¹⁵

He could have cut R&D, and didn’t. From 2001 to 2003 Apple spent $1,347 million on R&D. Over those same three years it earned $109 million in total. In its 2003 annual report Apple said that speeding up innovation was the best answer to a bad economy. Mac OS X, the iPod and the music store came out of that stretch.¹⁶

The company’s own cash paid for the Apple Stores: 65 of them between May 2001 and September 2003, for $290 million. In 2011 those stores and the ones that followed sold $14.1 billion.¹⁷

He also used the cash to lock in parts. In 2005 he prepaid $1.25 billion to five flash memory makers. In 2009 he did it again with $1 billion: $500 million to LG Display for screens and $500 million to Toshiba for flash memory.¹⁸

He bought almost no companies. In 14 years Apple paid $1.2 billion for acquisitions. They were small technology deals that ended up inside its products: Emagic (2002), maker of the Logic music software; P.A. Semi (2008) and Intrinsity (2010), both chip designers; Quattro Wireless (2010), in mobile advertising; and Siri (2010). The busiest year was 2010, with $638 million spread over several deals.¹⁹

Sales grew 18x over.

What Jobs didn’t do

From 1998 to 2011 Apple generated $90.1 billion in operating cash flow. Its cash pile grew by $80.1 billion. That is more than the whole company was worth on the stock market in 2006.²⁰

He never paid a dividend. The last one went out in fiscal 1996, before Steve came back.²¹

He bought back $217 million of stock in 14 years, and he did it backwards. Apple paid $15 a share in 1999 and $22.75 in 2000. In 2002 the stock closed the fiscal year at $7.36. Of every dollar a buyer paid for Apple, 82 cents was cash. That year the company bought back nothing.²²

Meanwhile the share count grew 32% from 2001 to 2011, because of the stock issued under employee plans.²³

And the cash earned less each year: 5.27% in 2007, 0.77% in 2011.²⁴

When Jobs called Buffett, one dollar in five of Apple’s market value was cash.

Final thoughts

So was Steve a good capital allocator? I think so, with one weak spot.

Inside the business he allocated well. He cleaned up the balance sheet, kept R&D whole through the bad years and paid for the stores out of Apple’s own cash. The weak spot was what he did with the money left over. In 2002, with the stock at $7.36 and 82 cents of cash behind every dollar of price, a cheap buyback sat in front of him and he passed. Eight years later he told Buffett himself that the stock was very undervalued, and he passed again.

Part of that was taxes. In 2010 about 60% of the cash sat in foreign subsidiaries, and bringing it home meant a tax bill.²⁵

He ignored Warren on buybacks. I still think Jobs was already a student of the Oracle of Omaha. Keeping cash on the balance sheet in case things go wrong is one of the lessons Buffett has taught us. Jobs followed it to the letter.

The difference is that at Berkshire the cash has two jobs, and Warren puts a size on the first one. For the bad days, Berkshire pledges to hold at least $10 billion and tends to keep at least $20 billion. Anything above that waits for a purchase: at the end of 2010 it had $38 billion, and Warren wrote that he wanted to put it to work.²⁶ Jobs held cash only for the bad days and for the freedom to take big risks.²⁷

In 2010 Apple had $51 billion, no debt, and $7.3 billion in operating expenses for the year.²⁸ The cash could have covered seven years of expenses without selling a single product.

Jobs never said how much was enough.

1. Warren Buffett, statement to CNBC after Jobs’s death, October 6, 2011. CNBC.

2. Warren Buffett on CNBC’s Squawk Box, February 27, 2012, as reported by MacTrast. The account of the call, including the quote, comes from this interview.

3. Cash, cash equivalents and marketable securities: $39.8 billion on December 26, 2009, and $51 billion on September 25, 2010. Apple Form 10-Q for the quarter ended December 26, 2009 and Form 10-K for fiscal 2010.

4. The deal valued BNSF at about $44 billion, including $10 billion of debt. Berkshire Hathaway and BNSF joint press release, November 3, 2009. Closing in February 2010: The Rational Walk.

5. Germany’s military expenditure in 2010 was an estimated $45.2 billion. SIPRI Yearbook 2011.

6. My calculation: $200 and $300 divided by 28, for the 7-for-1 split of 2014 and the 4-for-1 split of 2020.

7. Closing prices of $204.92 on October 21, 2009, and $300.14 on October 13, 2010. Yahoo Finance.

8. iPhone sales of $6.7 billion in fiscal 2008, $13 billion in 2009 and $25.2 billion in 2010. Apple Form 10-K for fiscal 2010.

9. Net income of $8.2 billion in fiscal 2009, $14 billion in 2010 and $25.9 billion in 2011. Apple Form 10-K for fiscal 2011.

10. Apple Form 10-K/A for fiscal 2009, which restated 2009 net income to $8.2 billion, and the October 2009 earnings release, which reported $5.7 billion under the old rule.

11. Medical leave announced January 14, 2009: CNN Money. Closing low of $78.20 on January 20, 2009: Yahoo Finance. Apple listed its dependence on key personnel, its CEO included, as a risk factor in its Form 10-K for fiscal 2009.

12. My calculation: market value at fiscal year-end 2010 minus cash ($216.8 billion), divided by net income for fiscal 2010 and fiscal 2011. Share count and net income from Apple’s Forms 10-K; closing price from Yahoo Finance.

13. Net loss, cash, long-term debt, the Microsoft investment and the debt ratings (B3 and Caa2 at Moody’s, B- and CCC at Standard & Poor’s) are in Apple’s Form 10-K for fiscal 1997. The filing calls the ratings “non-investment grade.”

14. Steve Jobs at the D8 conference, June 2010, as reported by Fortune.

15. Apple Forms 10-K for fiscal 1998, 1999 and 2004.

16. R&D of $430 million, $446 million and $471 million, and net income of -$25 million, $65 million and $69 million, in fiscal 2001, 2002 and 2003. Apple Form 10-K for fiscal 2003.

17. Apple had 8 stores at the end of fiscal 2001, 40 at the end of 2002 and 65 at the end of 2003, per its Form 10-K for fiscal 2003. The $290 million is retail capital spending for fiscal 2001 to 2003, from the Forms 10-K for those years. First two stores, May 19, 2001: Apple press release. Retail sales of $14.1 billion: Form 10-K for fiscal 2011.

18. Flash memory prepayments: Apple press release, November 21, 2005. LG Display and Toshiba prepayments: Apple Form 10-K/A for fiscal 2009.

19. Cash paid for acquisitions by fiscal year, from Apple’s cash flow statements: $10 million (1998), $52 million (2002), $220 million (2008), $638 million (2010) and $244 million (2011). Emagic is named in the Form 10-K for fiscal 2002. Apple’s filings do not name the deals from 2008 on; those names come from press reports compiled in Wikipedia’s list of Apple acquisitions.

20. My sums from the cash flow statements in Apple’s Forms 10-K for fiscal 1998 to 2011 (full list on EDGAR). Apple’s market value at the end of fiscal 2006 was $65.8 billion: closing price times shares outstanding.

21. Apple Form 10-K for fiscal 1997.

22. Buybacks under the plan approved in July 1999, from Apple’s Forms 10-K for fiscal 1999, 2000, 2003 and 2007. Prices are adjusted for the 2005 split, the same basis as the $200 in Buffett’s account. The 82 cents is cash and investments divided by market value at the end of fiscal 2002, my calculation.

23. Shares outstanding at fiscal year-end, adjusted for the 2005 split: 702 million in 2001 and 929 million in 2011. Apple Forms 10-K for fiscal 2001 and 2011.

24. Weighted average interest rate earned on cash and investments. Apple Forms 10-K for fiscal 2007 and 2011.

25. Foreign subsidiaries held $30.8 billion of the $51 billion. Apple Form 10-K for fiscal 2010.

26. Berkshire Hathaway 2010 shareholder letter. The $10 billion pledge excludes cash held at the regulated utility and railroad businesses.

27. Jobs at Apple’s shareholder meeting, February 25, 2010, as reported by Fortune and AppleInsider.

28. R&D of $1.8 billion plus selling, general and administrative expenses of $5.5 billion. Apple fiscal 2010 results.

    ---

    ---

    What do you think about Steve’s capital allocation skills? Was he a good capital allocator? Drop your comments!

    And if you liked this post, share the love!

Share

On the subject of economies of scale, I find chain stores quite interesting. Just think about it.

The concept of a chain store was a fascinating invention. You get this huge purchasing power—which means that you have lower merchandise costs. You get a whole bunch of little laboratories out there in which you can conduct experiments.

And you get specialization. If one little guy is trying to buy across twenty-seven different merchandise categories influenced by traveling salesmen, he’s going to make a lot of dumb decisions.

But if your buying is done in headquarters for a huge bunch of stores, you can get very bright people who know a lot about refrigerators and so forth to do the buying.

The reverse is demonstrated by the little store where one guy is doing all the buying. So there are huge purchasing advantages.

Some come from simple geometry.

If you’re building a great circular tank, obviously as you build it bigger, the amount of steel you use in the surface goes up with the square and the cubic volume goes up with the cube.

So as you increase the dimensions, you can hold a lot more volume per unit area of steel.

There are all kinds of things like that where the simple geometry—the simple reality—gives you an advantage of scale.

If this was worth your five minutes, subscribe to Ezra's Notes. And if you liked the post, share the love!

Share

Of every $100 it takes in, $60 comes from selling dough, cheese and boxes to the people who own the stores. Another $21 is royalties and fees, a cut of every pizza someone else sells. Only $8 comes from its own stores. The last $11 is advertising money that passes through its books and leaves.

That is why it’s such a good business. The franchisee pays for the store, the oven and the drivers. DPZ collects. With that model, EPS quadrupled between 2016 and 2025.

It is also why the company has a problem that doesn’t show up in its accounts. Domino’s doesn’t decide how many stores open. That call belongs to someone who puts about $400,000 of their own money on the line and runs the numbers. Those numbers look worse today than in 2019, and it shows. The company promised 1,100 new stores a year. It opened 775 in 2024 and 776 in 2025.

The market still prices Domino’s as if it will grow the way it used to. At $303/share, the stock assumes 11% annual growth for 5ys. Management promises 8%. The business has delivered 5%.

It’s a good business, and at this price it’s expensive. Below I explain both.

(Full research report at the end of this post)

The Thesis

DPZ is a tollbooth. It collects 5.5% of sales from 6,924 U.S. stores it doesn’t own, and it sells them their supplies too. 99% of franchisees renew. Cash comes in as fast as profit: operating cash flow beat net income in each of the last 10y. And for a decade it has been taking customers from its rivals.

The last decade won’t repeat. Almost half the growth in EPS came from lower taxes and buybacks funded with debt. Franchisees pay for the growth, and their stores earn less than they used to. Management promises more than it delivers. And in July 2027, Domino’s has to refinance $1.3B of debt at higher rates.

Will franchisees earn enough to keep opening stores? If the math works for them, Domino’s grows almost for free. If it doesn’t, Domino’s stops growing, however good its business is.

A toll on every pizza

There are 22,142 stores with the Domino’s logo around the world. The company owns 262 of them. Franchisees own the rest: they rent the space, hire the drivers and bake the pizzas. Domino’s supplies the brand, the app, the advertising and the dough, and charges for all of it.

The company reports three segments:

1. U.S. stores: royalties and technology fees from franchisees, the advertising fund and the 262 company-owned stores. It brings in 33% of revenue and 49% of profit.

2. Supply chain: 30 centers in the U.S. and Canada make the dough and deliver ingredients to more than 7,800 stores. It is 60% of revenue but keeps only about 11 cents of each dollar it sells, and Domino’s hands half of that profit back to franchisees. Buying from it is optional.

3. International franchise: 14,956 stores in more than 90 markets, each run by a master franchisee that pays about 3% of its sales. It is 7% of revenue and 24% of profit. With no cost of goods to speak of, 85 cents of every dollar ends up as profit.

Over 10y, all three grew at almost the same pace. Excluding the advertising fund, revenue rose 6.6% a year, and each segment’s profit grew between 8.5% and 9.3%.

Two crises tested the model. Between 2006 and 2009, operating income fell 11.5%. The stock still lost 81% during 2008, because a year earlier Domino’s had borrowed up to 7.6x its EBITDA. In 2022, inflation cut store profits by about 20%. Demand held up both times. What broke first was the balance sheet, and then the franchisee’s wallet.

That is where the model hits its limit. Domino’s grows when stores open, and someone else pays for the stores. From 2016 to 2019 the system added about 1,120 net stores a year. From 2022 to 2025, about 820. The average U.S. store now earns $166,000 a year, against $133,000 in 2016. That is 2.5% a year, over nine years in which Domino’s own operating income doubled.

Why Domino’s wins

Americans spend $43.4B a year on quick-service pizza, and the market grows 2.5% a year. The chains say they compete on quality, service, technology and price. In practice they compete on two things: the price of this week’s deal and how long the pizza takes to arrive. Nobody has pricing power. Customers lose nothing by switching chains.

Even so, market share has moved in one direction. Domino’s holds 23.3% of the market. Between 2015 and 2023 it gained 9.4 points, while Pizza Hut, Papa Johns and Little Caesars together lost 3.4 most of what Domino’s gains comes out of its big rivals.

Size explains it. DPZ advertising fund raised $559M in the U.S. in 2025. Pizza Hut’s raised $360M worldwide. One supply chain buys the cheese and flour for the entire system. So the same deal leaves more money in a Domino’s store, which sells $1.39M a year, than in a Pizza Hut, which sells $0.81M. CEO Russell Weiner put it this way in February: “We have profit power.” Domino’s can sell cheap and its franchisees still make money.

What it doesn’t have is a captive customer. Independent pizzerias and regional chains still take about 39% of delivery. And the barrier that held them back, the cost of running their own drivers, fell when Uber Eats and DoorDash arrived. Domino’s refused to sell on those apps for years. It joined them between 2023 and 2025.

Meanwhile, the rivals are retreating. Pizza Hut closed 250 U.S. stores in 2025, and Yum! Brands sold it to a private equity firm in September 2026. Papa Johns closed 101 restaurants in North America in the first half of 2026 and suspended its dividend.

Management delivers cash, not growth

Domino’s management promises more growth than it delivers, and when it misses, it moves the goalposts.

In January 2019, management talked about 25,000 stores and $25B in sales by 2025. It ended that year with 22,142 stores and $20.1B. In December 2023 it unveiled a new plan through 2028: 1,100 net new stores a year, sales growing 7% and operating income growing 8%. The store target lasted seven months. In July 2024 the company suspended it “temporarily,” and the stock fell 13.6% that day. Nineteen months later it was still suspended. The 2026 guidance was cut two months after it was issued.

Of 15 growth targets for 2024 and 2025, the company hit three. It hit all four of its cost and cash targets. It delivers what it controls and misses what depends on franchisees wanting to open. That’s why I value Domino’s on what it has delivered, not on what it promises.

The pay plan doesn’t help. The CEO earned $10.7M in 2025, and only 9% of it was fixed. The bonus and most of the stock awards depend on an adjusted EBITDA figure that the compensation committee defines itself. FCF, ROIC, per-share results and store openings all go unmeasured. When results fell short in 2022, the committee changed the rules. In January 2023 it lowered the bonus target a month after setting it, and it redesigned the performance shares. The first cycle under the new rules paid out at 121.8%, after the two before it paid 98.5% and 59.6%. And the internal targets that trigger pay ask for less than the company promises in public.

Management doesn’t own much of the company either. Directors and executives hold 0.89% of Domino’s, down from 3.57% at the end of 2016. The CEO’s own shares are worth about $8.9M, less than he was paid in 2025. Since 2016 there has been exactly one open-market purchase by an insider. The real owner sits outside: Berkshire Hathaway holds 9.96%.

As for the cash, the company reinvests little, earns a lot on it, and returns the rest. The problem is the price it pays for buybacks. Between 2016 and 2019 it bought back 12.8M shares at an average of $207, and those shares are worth 46% more today than they cost. From 2020 to 2025 it paid an average of $421. Those shares cost $2.87B and are worth about $2.07B today, 28% less. Nobody at the company tells the two apart in public, and neither does the pay formula.

Lots of cash, little capital, plenty of debt

Between 2016 and 2025, Domino’s sales excluding advertising grew 6.6% a year. FCF/share, after stock-based compensation (SBC), grew 15.1% a year, from $5.17 to $18.31.

Not all of that gap came from the business. Operating income grew 8.6% a year and accounts for 53% of the growth in EPS. The rest came from three places: 31% fewer shares outstanding, a tax rate that fell from 37.7% to 21.9%, and interest costs that grew more slowly than profit. The tax cut won’t repeat. And part of the buybacks was paid for with debt.

What makes DPZ unusual is how little capital it needs. Franchisees pay for the stores, so Domino’s is left with distribution centers, software and working capital: $533M at the end of 2025. On that base it generated $792M of operating cash flow. ROIC was 140% in 2025 and averaged 154% over 10y.

Then there’s the debt. Domino’s owes $4.8B, 4.4x EBITDA. The notes carry fixed rates and are backed by royalties and the brand. Relative to cash flow the load hasn’t grown: since 2016 debt has risen 2.2x and FCF 2.4x.

What I’m watching is the cost of rolling it over. In July 2027, $1.3B comes due, about a quarter of all its debt, at an average rate of 4.2%. The notes Domino’s issued in September 2025 already priced at 4.93% and 5.22%, and the five-year Treasury yield has since climbed from about 3.7% to 5.06%. At the same spread it got in 2025, Domino’s would pay about 6.3%. If all five series maturing between 2027 and 2031 moved to that rate, interest would rise by about $103M a year, more than half of what Domino’s paid in 2025.

Do I trust the numbers? Yes. Operating cash flow beat net income in all ten years, the Beneish M-Score averages −2.63, well below the −1.78 line that flags possible manipulation, and there are no restatements and no acquisitions to hide anything in. The one caveat is 2025. Cash flow got a lift that year because Domino’s paid its suppliers later. Strip that out and FCF was $574 million, which is the figure I use to value the company.

What it’s worth

At $302.87/share, Domino’s has a market cap of ~$10B, roughly what its U.S. stores ring up in a single year. Add the debt and the enterprise value is $14.7B.

First, where it trades. Against five other franchisors, Domino’s sits in the middle: 15.2x operating income, versus 16 to 19x for McDonald’s , Yum! Brands and Restaurant Brands International . On cash flow the gap is wider, 19x against 23 to 25, because Domino’s needs less investment. At the peer median, the stock would be worth somewhere between $328 and $403. That tells me it isn’t expensive relative to its peers. It doesn’t tell me what it’s worth.

For that, I project its cash flow ten years out and discount it back at 10%. My rule is simple: more risk, higher rate. The business is stable. The balance sheet less so, and growth depends on someone else’s wallet.

I ran three stories:

• Bear case, $137/share: franchisees open fewer stores and cash flow grows 3% a year.

• Base case, $205: Domino’s keeps growing 5%, what it has delivered since 2021, starting from the $574M of 2025.

• Bull case, $284: management hits its 8% target for five years.

Management’s target sits in the bull case rather than the base case because of the track record: 3 out of 15.

Then I ran the math backwards. For the stock to be worth $303, DPZ would need to grow 11% a year for 5y. That’s more than it promises and twice what it delivers. If my base case is right, a buyer today earns 8.4% a year. Not a bad result, but below the 10% I require.

Conclusion

Domino’s stays on my watchlist. At $303, the price leaves no margin of safety.

It’s a good business. It takes a cut of every pizza while putting up almost no capital, turns its profit into cash, and has spent a decade taking customers from rivals that are now closing stores. What changed is the pace. For ten years EPS compounded at 17%, and almost half of that came from taxes and debt. What’s left is a business growing 5% a year, run by a management team that promises 8%.

For a business of this quality I want a margin of safety of at least 20%, which means buying at least 20% below what it’s worth. Against my value of $205, I think the sensible price to pay would be $164 a share or less. That’s a suggestion based on this analysis, not investment advice.

At $164, DPZ would trade at about 10x EBIT. That price may never come. I’d rather miss a good company than pay for growth that isn’t in the numbers yet.

DOWNLOAD THE FULL REPORT

Dominos' Pizza Ezra's Notes Full Research Report

7.24MB ∙ PDF file

Download

Download

DISCLAIMER

This analysis is for educational and informational purposes only. It is not investment advice or a recommendation to buy or sell any security.

Do your own research or consult an adviser before making any investment decision.

I have no position in Domino’s Pizza (DPZ).

Data as of the October 6, 2026 close.

Would you buy Domino’s at $303, or would you wait for something closer to $164?

If you think I got something wrong, say so in the comments.

And if you liked this post, share the love!

Share

AUDIENCE MEMBER: Mr. Buffett, my name is Aaron Morris (PH). I’m from California.

What I wanted to know was how you think about how large a position you’re willing to take in a given security, both in your case, where you have new cash coming in that you can invest, and in the case of an investor, where they have a fixed amount of capital, and they’re trying to decide what’s the most (inaudible) security that they really love?

WARREN BUFFETT: Well, Charlie and I have — probably at our present size, we will never find anything that we get as much money into as we want. Don’t you think that’s probably true, Charlie? If we really like it?

CHARLIE MUNGER: Yeah, I think that’s quite likely.

WARREN BUFFETT: Yeah, so we will probably never hit the limit. We would love to. We’d love to find something we felt that strongly about, and occasionally we do.

But we won’t get as much money into it as we would wish, or as if we were running a million dollars of our own money or some number like that, so —

We are willing to put a lot of money into a single security. When I ran the partnership, the limit I got to was about 40 percent in a single stock. I think Charlie, when you ran your partnership, you had more than 40 percent in —

CHARLIE MUNGER: Sure. (Laughter)

WARREN BUFFETT: And we would do the same thing if we were running smaller partnerships, or our own capital were smaller and we were running that ourselves.

Because, no, we’re not going to do that unless we think we understand the business very well, and we think the nature of the business, what we’re paying for it, the people running it, and all of that lead up to virtually no risk, and —

But you find those things, occasionally. And we would put — assuming it were that much more attractive than the second, and third, and fourth choices — we would put a big percentage of our net worth in it.

We only advise you to do that — well, we probably don’t advise you to do it all, maybe — but we would only advise you to do it, if you’re doing it based on your conclusions about — your own ideas of value, and something that you really feel you know enough to buy the whole business, if your funds were sufficient, and it was being offered to you. You ought to really understand the business.

But people do that all the time, incidentally, in private businesses, which have got terrible prospects. I mean, they buy dry cleaning establishments, or filling stations, or whatever, and they put very high —franchises of some kind — they put a very high percentage of their net worth into something — a business that’s very risky, basically. I mean, it —

People put all their money in a farm, you know. It’s a business. It’s subject to all kinds of business risk.

So it’s not crazy, if you understand the business well, and if the price is sufficiently attractive, to put a very significant percentage of your net worth in. If you don’t understand businesses, then you’re better off diversifying and fairly widely diversifying.

CHARLIE: Berkshire has a substantial shareholder whose father accumulated the original position, and when he died he left a very large estate, practically all of which was in two securities, Berkshire and one other outstanding company.

A bank was co-trustee. And the bank trust officer said you’ve got to diversify this. And, you know, it was a very large estate.

And the young man who was co-trustee with the bank said, “Well,” he says, “you know, if my father believed the way you do, he might have been a trust officer in a bank instead of — (laughter) — leaving this large estate.” (Applause)

And that young man holds the Berkshire to this day. And I suppose the bank is still giving the same advice. (Laughter)

If this was worth your five minutes, subscribe to Ezra's Notes. And if you know someone who diversifies out of habit instead of conviction, share it with them.

Share Ezra's Notes

Wix is the company that lets your barber, or your dog walker, build a real website without touching a line of code. Pick a template, drag some blocks, publish. That’s been the pitch since 2006.

It’s not the whole pitch anymore. Somewhere in the last seven years, “build a website” became “run your business from one screen” — payments, bookings, a domain, an app store. Whether that shift actually protected Wix is most of what this post is about.

But before we get to the breakdown, a word from today's sponsor: Ezra's Notes.

The Subscriber Slowdown

Vinall’s opening line on growth: revenue went from $757 million in 2019 to almost $2 billion in 2025 — “impressive,” in his word — but growth “lagged” his own expectations, mainly because subscriber acquisition slowed.

I assumed he meant premium subscriptions. So I pulled both numbers Wix reports — registered (free) users and premium (paying) subscriptions — and lined them up.

They diverge completely. Registered users grew 84% over the period. Premium subscriptions didn’t keep pace — and starting in 2023, the paying base shrank in absolute terms, not just relative to the free base.

Wix has never published a free-to-paid conversion rate. So I built one: premium subscriptions divided by registered users, year by year. It fell every single year of the seven-year period, no exceptions — from 2.72% to 2.01%, a 26% relative decline.

That number changes the story. Rob frames this as something that went wrong late — 2024, 2025. The filings say otherwise: the conversion engine had been leaking since the start. The last two years weren’t a break from the trend. They were the trend, finally visible in the headline subscriber count.

His original 2019 thesis had two separate growth engines packed into one sentence: “expanding website volume and higher revenue per user as clients integrated e-commerce functionality.” The data splits them cleanly. The first engine — website volume, i.e. paying subscribers — never delivered. The second did, and then some.

Business Solutions — Wix Payments, the app market, domains — nearly doubled its share of revenue, from 15.3% to 29.3%, over the same years conversion was falling.

The mechanism is simple: Wix Payments takes a cut of transaction volume, and that volume comes from the whole ecosystem — including free users selling through their sites — not from premium subscription counts. The company didn’t grow the way Vinall originally underwrote. It found a second engine and let that one carry the business instead.

Free Cash Flow

The second plank of the 2019 thesis: FCF margins would expand toward 30%, driven by “natural operating leverage and declining customer acquisition costs.” Wix did report a 30% margin in 2025. Two things make that number less clean than it looks.

First, the timing. FCF margin fell every year from 2019 to 2022 — from double digits into negative territory — until activist firm Starboard Value disclosed a stake in September 2022 with one explicit ask: fix the margins and improve capital allocation. The recovery that followed tracks the activist’s entry almost exactly, not a smooth internal curve.

Second, the definition. Vinall’s own words: management’s free cash flow figure “excluded stock-based compensation” — and once you add it back, the margin isn’t 30%. It’s closer to half that.

Why the gap reopened in 2025: Base44, Wix’s push into AI-assisted coding. On the Q4 2025 call, management said the quiet part out loud — they’re no longer optimizing gross margin, the thing Starboard asked for in 2022. They’re optimizing dollar profit.

One thing did work as advertised. Customer acquisition cost, measured as sales and marketing spend over revenue, fell from 44.3% in 2020 to 24.2% in 2024 — nearly cut in half.

The Moat

Original thesis: the moat would deepen as Wix evolved “from a simple content repository into an operating system,” raising switching costs for small business customers. Vinall’s post-mortem read: it did widen — but he’s no longer certain it holds, given AI, and he’s chosen to watch from the sidelines rather than hold an opinion with money behind it.

He’s right that it widened. Wix layered payments, bookings, a domain registrar, and an app marketplace on top of the original site builder. Every product a customer adopts makes the whole stack harder to leave — payments, calendar, and website all living on one platform raises the cost of migrating any single piece.

But Wix’s own 2025’ 20-F names the threat directly, in language a company has every incentive to soften, not sharpen:

Wix’s response wasn’t to defend the castle. It was to buy a stake in the army outside the walls. Base44 — the AI “vibe coding” startup the company acquired in 2025 — competes directly with Claude Code, Cursor, and Replit, the exact tools that could make Wix’s specialist software obsolete. If you can’t beat them, buy them.

Management

This is where the thesis breaks down hardest.

Rob’s original bet: a founder-led team with an owner’s mindset. His 2026 verdict: dedicated, yes — nobody left in seven years — but “growth rather than owner-oriented.” I went looking for the numbers behind that sentence.

Ownership. After nearly two decades running the company, CEO Avishai Abrahami owns 1.7% of Wix in actual shares — not options, not RSUs, shares. The full management team, combined, owns 7.2%.

Compensation. Wix’s own Compensation Policy — unchanged since November 2023 — requires at least 75% of the annual bonus to rest on measurable criteria. The policy names three: collections, revenue and pre-tax profits, and user counts. Not one mention of ROIC, or return on capital, or anything resembling long-term value creation. Vinall’s line lands exactly where the policy points: “It would have helped if they had not used a flawed free cash flow metric to track profitability.”

Capital allocation. Between 2021 and 2025, Wix spent $1.6 billion buying back its own shares. Over that same stretch, diluted shares outstanding didn’t fall — they climbed, peaking near 60 million in 2024, the same year as the second-largest buyback of the period.

A buyback built to create value has two ingredients: cash, and a management team that believes the stock is cheap. Wix had the first. There’s no record of the second — no year in this period where a repurchase program launched with a stated view that shares were undervalued. What the numbers show instead is a repurchase sized almost exactly to that year’s stock-based compensation, quarter after quarter — not “we think this is cheap,” but “how much did we dilute this year? Buy back roughly that much.”

Final Thoughts

Wix it’s just a company that chased the next growth engine whenever the first one slowed, run by a team with almost no skin in the game, paid to hit growth-vanity-metrics targets rather than long-term value targets — sitting behind a moat that's real, moderately deep, and, by the company's own admission, under a threat it decided to acquire rather than fight.

What would you have flagged that I missed? Drop it in the comments — I read all of them.

If this changed how you'll read the next shareholder letter that crosses your desk, subscribe to Ezra's Notes, and send it to the one person you know who'd argue with it.

Share

On March 17, 2009, eight days after the market's crisis low, Seth Klarman joined George Athanassakos' value investing class at Ivey Business School by video. I hadn't gone back to this talk in over five years — rereading it now, the lines that hold up are the ones on risk and psychology, not the market calls. These are the ones I kept underlining.

On temperament

On risk

On relative performance

On who’s on the other side of the trade

On the worst case

On valuation

On concentration

On selling

On integrity

If this was worth your five minutes, subscribe to Ezra's Notes. And if you know someone who'd keep these lines on their desk, share it with them.

Share Ezra's Notes

Today on Ezra’s Notes, I’m launching a new feature called Conversations: short interviews with investors, built around questions that lean more toward who they are than what they’ve achieved as investors.

The idea is to learn how their mind works, and what’s shaped how they approach investing in the first place.

For the first edition: a conversation with Mauricio Heck, aka

Mauricio’s been a close friend for 7 years now, since my MOI Global days. We’ve talked about everything from investing and digital business models to music.

This first edition means a lot to me.

But before we get to the interview, a word from today's sponsor: Ezra's Notes.

Favorite band or artist, and what does it say about how you think?

I don’t know if I have a single favorite of all time. If I had to choose, I would probably say The Beatles or U2, along with the great classic rock bands.

These days, however, I listen to a lot of Hans Zimmer. His scores often begin with a simple theme and build into something vast, layered, and emotionally powerful. Interstellar is a great example, as are Inception and Christopher Nolan’s Batman trilogy.

That way of creating resonates with how I think: start from first principles, find the simple idea underneath everything, and keep iterating until it becomes something much richer.

Last book you read that had nothing to do with investing?

Almost everything I read these days eventually connects to investing or to themes I explore in my newsletter, especially the broader skills that help someone become a better investor.

The closest answer right now is Atlas Shrugged by Ayn Rand, which I am rereading after many years. I consider it a masterpiece, although it certainly demands a serious time commitment.

Coffee, tea, or something else, and where do you do your best thinking?

Definitely yerba mate. It’s the perfect middle ground between tea and coffee. It gives me more lift than tea without feeling as strong as coffee.

It’s a South American drink prepared in a gourd, and for me it’s the ideal companion for intellectual work.

I do my best thinking in the morning. I used to be a late night person, but having a child has a way of turning you into a morning person.

A daily habit you would never give up, no matter how busy you get?

My structured morning routine. It’s the foundation for everything else I do. I start with exercise, including stretching and push-ups, followed by meditation and a cold shower. I try to protect that routine every day because it sets the tone for everything that follows.

Where do your best ideas come from, a walk, a shower, or a specific chair?

Meditation, surprisingly. I meditate for 15 to 20 minutes before bed and again early in the morning, and those are often my most creative moments.

Reading also sparks many ideas, so I always keep a notebook nearby and write them down as they come. Most ideas begin as a small observation in the middle of something else.

A skill you are proud of that has nothing to do with your job?

Sailing. I grew up in a coastal town with easy access to the sea and sailboats. I started sailing Optimist dinghies as a child, moved through several classes, and eventually competed in sailboat races. I used to be quite good at it, and it remains something I love.

What is the most contrarian opinion you hold that has nothing to do with markets?

I don’t know whether it qualifies as contrarian, but I am a committed believer in Carol Dweck’s growth mindset. I know aspects of her work have been debated, but the underlying idea has guided me throughout my life: abilities are not fixed, and we can develop them through learning, better strategies, and sustained effort. That mindset has helped me enormously over the years, and I think it is one of the most useful principles a person can carry through life.

What is a completely non-financial skill that has made you a better investor?

Reading widely, guided by curiosity. Curiosity has driven my entire investing career. I want to understand how companies work, how the world works, why trends emerge, and what might happen next.

The habit of following questions wherever they lead, combined with a genuine love of reading, has probably done more than anything else to make me a better investor.

If you were not in investing, what would you be doing?

I would probably do something outdoors, perhaps work as a nature guide or make documentaries in remote places. I love being outside, especially in places far from crowds, and I seriously considered that kind of career many years ago.

If you could have dinner with anyone, alive or dead, who is not an investor?

Messi or Maradona would be tempting choices, especially as an Argentine. But if I had to choose one person, it would be Leonardo da Vinci, with Benjamin Franklin a close second.

I am fascinated by people whose curiosity crossed many disciplines. They wanted to understand the world around them and refused to confine themselves to a single field. That is the kind of mind I would love to spend an evening exploring.

One thing you are terrible at that everyone assumes you are good at?

Math. People often assume that investors are naturally gifted with numbers. I am comfortable with the math my work requires, but it’s not where my strongest edge lies.

I gravitate much more toward the qualitative side of investing: mental models, incentives, understanding people, and connecting ideas across different fields.

If Mauricio’s answers surprised you, forward this to someone who'd enjoy meeting them too — and drop your favorite answer in the comments.

Share

Would you change your wife every year?

That was Chris Hohn’s answer when someone asked why he doesn’t rotate his portfolio more often.

Chris is the founder of TCI Fund Management. Over more than two decades, he’s compounded capital at 18% annually — roughly 9 points ahead of the market.

His philosophy: invest in quality companies, run a concentrated portfolio, and hold positions for years — some, like Moody’s, for more than a decade.

And like every great investor I’ve studied, he’s taught me a few things worth passing on.

First, eliminate

Most investors go looking for good ideas. Chris does the opposite: he eliminates almost everything before he even starts.

Banks. Airlines. Telecom. Media. Insurance. Commodities. Utilities. Asset managers. Oil & gas. Retail.

Why so ruthless? In his own words:

His argument is straightforward: in competitive industries, profits tend toward zero over time. It doesn’t matter how fast the sector grows.

Growth without barriers is worthless.

After that filter, Chris considers about 200 companies in the entire market investable — high quality enough to own. Two hundred, out of ~58,000 publicly listed companies.

This reminded me of Rupal Bhansali, a portfolio manager at Ariel Investments. She arrives at the same place, but through process rather than industry. She flips the analyst’s default posture: every company starts out rejected, not selected, and has to prove it belongs.

And both share the same rule number one:

Moats, moats, and moats

Once inside that narrowed universe, Chris has one obsession: what protects this business from competition? Not just any moat — durable ones. And ideally, more than one.

Irreplaceable physical assets

Airports, toll roads, cell towers, railroads. Nobody is going to build a second airport in Madrid. Nobody is going to lay a second set of tracks right next to the existing ones. These are natural monopolies.

This connects to something I wrote a while back about Grupo Aeroportuario del Centro Norte (OMA), an airport operator in Mexico. Nobody would build a second airport in Monterrey. OMA doesn’t need to reinvent air travel. It just executes within lines that were already drawn. This is a low-entropy business. Chris, without using that term, is looking for exactly the same thing.

Advanced intellectual property

His favorite example here: jet engines — and it shows. GE Aerospace , at roughly 33% of the portfolio, is his single largest position. There are only two players in narrow-body engines and two in wide-body. There hasn't been a new entrant in over fifty years. The engines run at temperatures so high that metals melt. Thousands of complex parts. It's simply too hard to replicate.

Network effects

Visa , his second-largest position, is his textbook case. Every new customer and every new bank that joins the network makes it more valuable for everyone else — and harder to displace.

Switching costs

Microsoft used to be the classic example in Chris's portfolio — until it wasn't.* With Office, Teams, and Azure all under one roof, switching costs are high.

Installed base

Back to jet engines: once the engine is on the wing, the airline depends on you for parts and maintenance for decades. It’s a recurring revenue stream that doesn’t require winning the sale over and over again.

The question Chris asks about every investment isn’t “can this company grow?” It’s “can someone else come in and compete with it in twenty years?”

If the answer is clearly no — invest. If there’s any doubt — pass.

The most common mistake: confusing growth with value

If there’s one thing the consensus loves, it’s growth — and it mistakes that growth for value. Growth only has value when it comes with barriers. Without them, growth benefits customers and competitors — not shareholders.

For Chris, growth itself isn’t the problem. What he wants are what he calls “super-companies”: ones that can raise prices above inflation, year after year. Pricing power at its purest.

Moody’s , his third-largest position, is the clearest example. It’s grown revenue at 10% a year for a hundred years. The company has pricing power, a captive customer base that can’t function without it, and virtually no real competition.

Time as a competitive advantage

Hohn’s average holding period is 8 years. That’s not just patience as a virtue. It’s a structural rule for the portfolio.

Why? Because the market prices companies looking 2 or 3 years out. If you’ve found a business that will still be excellent in 20 years, the market gets it wrong. The gap between what the market sees and what you see — that’s your edge.

But there’s a catch: holding onto a quality company only works if the fundamentals were right to begin with — if the moat is real and durable.

Final Thoughts

Chris’s investment philosophy comes down to three questions:

1. Does this company have real, durable moats?

2. Can it raise prices above inflation?

3. Can you hold it for 10 years with conviction?

And it sums it all up in one line:

And the best way to avoid permanent loss is to stay out of industries where competition can destroy you, not overpay for growth without a moat, and not sell what’s genuinely great just because the market is asking you to.

Find the right company. Understand it deeply. And don’t change it every year.

Chris’ Checklist

• Avoid banks, airlines, telecom, media, insurance, commodities, utilities, retail, oil & gas, and asset managers.

• Find a real, durable moat: irreplaceable physical assets, hard-to-replicate IP, network effects, switching costs, or an installed base.

• Make sure it holds up against competition, technology, and 20 years of time.

• Confirm pricing power, year after year.

• Hold it for the long-term, with conviction.

* TCI cut its Microsoft position from 10% to 1% of the portfolio in Q1 2026, then exited entirely in Q2. Chris's reasoning: "Rapid progress in AI creates uncertainty around Microsoft's future competitive position. We are primarily concerned about the Microsoft Office productivity software franchise: AI could change established workflows and lead to new work platforms. In addition, we see some risks to [cloud service] Azure." (Source: en.oninvest.com)

Download this post in PDF

What Chris Hohn Taught Me About Rejecting 99

197KB ∙ PDF file

Download

Download

Which of your current positions would survive Chris's filter — and which wouldn't? Tell me in the comments — I read them all.

If Chris's checklist changed how you'll look at your next idea, the best way to support Ezra's Notes costs nothing — forward this to one person who'd actually use it.

Share

You instantly recognize them. Not just because of the color, but because of that iconic emblem known as the Cavallino Rampante (Prancing Horse).

Yes, I’m talking about Ferrari.

“Oh, a luxury car company,” you might say.

But not exactly.

At its core, Ferrari is a company that manages one of the most desirable brands in the world. It’s not just another automaker.

Yes, the cars are the product.

But the business is something else entirely.

Ferrari is a luxury house, more comparable to LVMH than to a traditional car manufacturer.

What Ferrari Really Does

Ferrari designs, manufactures, and sells high-performance sports cars.

But unlike most automakers, it does not compete on volume, price, or market share.

It competes on something far scarcer: desirability.

And that changes everything.

The Business Model

To survive in a highly competitive industry, most car manufacturers try to produce and sell more vehicles every year.

Ferrari is different.

It operates under a simple idea:

Always produce one fewer car than the market demands.

That decision has several consequences:

• Demand remains higher than supply

• Ferrari doesn’t need discounts to sell cars

• It can raise prices without hurting demand

“Hold your horses, Ezra… are you telling me Ferrari competes by making fewer cars?”

Yes.

Because scarcity is not a limitation of the business.

It is the product.

Ferrari doesn’t look for customers.

It builds collectors.

Its business model doesn’t end when the car is sold.

That’s where it begins.

In 2025, 84% of Ferrari’s new cars were purchased by existing Ferrari owners. And more than half of those were sold to people who already owned multiple Ferraris.

Over time, the customer stops buying a car.

They begin building a collection.

They become part of a community.

Ferrari creates a long-term relationship through:

• Extreme personalization

• Access to limited models

• Preferential treatment

• A closed community

The Invisible Asset: Resale Value

Beyond the community, one of the most important parts of Ferrari’s model is how aggressively it protects resale value.

That completely changes the buying decision.

If a customer believes their Ferrari will hold — or even increase — its value over time, the real cost of ownership becomes much lower.

In some cases, even negative.

Why does that matter?

Because it makes buying the next Ferrari much easier.

The secondary market is not a side effect of the business.

It is part of the business.

Limited production, personalization, community, and resale value are not independent decisions.

They are designed to reinforce one another.

Ferrari does not sell individual cars.

It builds a system that compounds value over time.

The Role of Dealerships

Ferrari does not own its dealerships.

But it controls them rigorously.

They are not treated as simple distributors, but as extensions of the brand itself.

Ferrari uses dealerships to:

• Control the customer experience

• Manage vehicle allocation

• Maintain exclusivity

• Protect resale value

• Serve as community touchpoints

The dealership does not simply sell cars.

It filters, nurtures, and cultivates the customer.

A Self-Reinforcing System

Ferrari’s business is not linear.

It is a system.

Racing creates visibility.

Visibility creates desirability.

Desirability enables premium pricing.

Premium pricing generates profits.

Profits are reinvested into technology and racing.

And the cycle repeats itself.

Competitive Advantage

At first glance, Ferrari’s model may appear replicable.

It isn’t.

The company is protected by several powerful shields:

• More than 75 years of history

• Continuous presence in Formula 1

• An extremely loyal customer base

• Tight control over distribution

None of this is built quickly.

And none of it can simply be bought.

Final Thoughts

To close this profile, I’ll leave you with a comment from Guy Spier, CEO of Aquamarine Capital, during an interview with MOI Global:

If you’d like to keep reading pieces like this one, I’d be glad to have you as a subscriber.

Ezra’s Notes is free, and it helps me write more essays worth your time.

Subscribe now

If you missed Part I, I’ll leave the link below so you can read it before continuing.

But structure can emerge from other forces as well.

Not all order comes from repetition or restraint.

Sometimes it comes from rules.

Sometimes from ritual.

Sometimes from geography.

The companies in this second part operate in very different industries. Airports. Spirits. Alpine railways. Yet each is anchored to something that moves slowly, or not at all.

One operates inside contracts written decades in advance.

Another aligns with habits that outlive marketing cycles.

The third controls access to a mountain that cannot be relocated.

Low entropy takes many forms.

These are three more.

Grupo Aeroportuario del Centro Norte (OMA) — A business that operates inside fixed lines

The runway does not move.

Flights are delayed.

Passengers complain.

Airlines change routes.

But the runway stays exactly where it is.

That stillness is not just physical. It is contractual.

OMA operates under long-term concessions granted by the Mexican government. These agreements do not merely allow the company to manage airports; they define the boundaries of the business decades in advance. The network is fixed. The geography is fixed. The framework is fixed.

In most industries, companies fight to expand their territory. OMA begins with it already assigned.

That constraint creates order.

At the heart of the model lies a simple separation. On one side, aeronautical services: landing fees, passenger charges, use of gates and security infrastructure. These are regulated. Prices are capped. Returns are negotiated through a maximum tariff system.

On the other side, commercial activity: restaurants, parking lots, advertising, hotels, logistics parks. These are largely unregulated. Here, the company has flexibility.

The result is a deliberate balance. The core utility is stable but capped. The surrounding ecosystem is flexible but anchored to passenger flow.

The airport becomes less like a transportation business and more like a regulated mall attached to a runway.

Every few years, another stabilizer enters the system: the Master Development Program. Investments are not impulsive. They are negotiated in advance with the government. Expansion plans are tied to tariff adjustments. Capital spending and future revenue are linked in a structured loop.

Growth, in other words, follows a script.

This matters. Because in many capital-intensive industries, volatility comes from mistimed expansion. Too much runway. Too little traffic. OMA’s structure reduces that risk by forcing discipline into the investment cycle itself.

Even internally, the company avoids fragmentation. Its thirteen airports are not run as isolated profit centers. Profitable hubs help sustain smaller locations through a solidarity mechanism. Administrative functions are centralized. Standards are shared. The network behaves as one organism, not thirteen separate ones.

Entropy often enters when units compete with each other or drift apart. OMA’s structure prevents that drift.

Airports look dynamic from the outside. Planes taxi. Screens flicker. People rush through terminals. But beneath the motion lies a system built around permanence—steel, concrete, contracts, and regulatory formulas that move slowly.

That is the hidden advantage.

OMA does not need to reinvent air travel. It does not need to persuade customers to adopt a new habit. It does not need to redesign its core product every season. Its primary task is operational execution within clearly defined limits.

Some businesses generate complexity by chasing expansion.

Others reduce complexity by accepting their boundaries.

OMA belongs to the second group.

Its low entropy does not come from innovation or cultural ritual. It comes from operating inside fixed lines—and allowing those lines to do most of the stabilizing work.

When the rules are written decades in advance, the system has fewer ways to surprise you.

Diageo — A business built on ritual

The bottle shape does not change.

The label remains familiar.

The liquid tastes the same.

The ritual repeats.

A wedding toast.

A holiday dinner.

A quiet drink after work.

Diageo’s advantage begins there.

Unlike many consumer companies, Diageo does not sell novelty. It sells continuity. Its portfolio is built around what it calls “icons”—brands that have survived decades, sometimes centuries, without needing reinvention. Scotch, tequila, gin, stout. Categories that are less fashion than tradition.

That distinction matters.

Most consumer businesses decay when tastes shift. They chase trends, launch extensions, refresh packaging, and experiment constantly to stay relevant. Diageo’s core brands do not need to chase. They occupy rituals that already exist.

The company’s strategy reflects that understanding. Rather than focusing on volume growth, Diageo leans into premiumisation—the idea that consumers may drink less, but drink better. As incomes rise, the trade-up effect supports value growth even when total consumption stabilizes.

The system is not built on acceleration. It is built on upgrading.

Behind the label lies another stabilizer: time. In spirits, maturation is not optional. Scotch ages for years. Sometimes decades. The company’s vast inventory of maturing stock represents revenue secured far in advance. Decisions made today will only be realized years from now.

Few businesses operate with such embedded patience.

Supply chains stretch from grain fields and agave farms to distilleries and warehouses filled with aging barrels. Yet despite the scale, the model remains coherent. Diageo centralizes operational processes globally while allowing local market teams to focus on execution. The structure absorbs complexity so the brands do not have to.

Entropy often enters consumer goods through fragmentation—too many SKUs, too many campaigns, too many experiments. Diageo’s discipline lies in protecting the core. The icons remain central. Innovation happens around them, not in place of them.

Even shifting social norms are incorporated rather than resisted. As moderation trends rise, the company expands into non-alcoholic variants, preserving the ritual even when the alcohol content changes. The habit survives. The brand remains.

That is the deeper layer of stability.

Alcohol consumption may fluctuate with economics and culture, but human rituals move slowly. Celebrations persist. Social bonding persists. Identity expressed through brand choice persists.

Diageo does not need to create those behaviors. It aligns with them.

Some businesses depend on constant novelty to stay alive.

Others endure because they anchor themselves to traditions that outlive any marketing cycle.

Diageo belongs to the second group.

Its low entropy comes not from avoiding change, but from embedding itself in habits that resist rapid change. When a business is tied to ritual, time works with it rather than against it.

And rituals, unlike trends, are patient.

Jungfraubahn Holding — A business anchored to a mountain

The mountain is not moving.

Snow melts.

Tourists arrive and leave.

Currencies fluctuate.

But the mountain stays exactly where it has always been.

Jungfraubahn’s advantage begins with that fact.

The company operates railways and cableways that lead to one of the most iconic high-altitude destinations in the Swiss Alps: Jungfraujoch — “Top of Europe.” It is not simply transportation infrastructure. It is access to a fixed point in geography, carved into rock more than a century ago.

Unlike most leisure businesses, Jungfraubahn does not invent demand. It channels it.

The model is built around control of the entire service chain. A visitor does not just buy a train ticket. They purchase a journey: transport, restaurants, retail, observation decks, curated experiences. By integrating the elements along the route, the company reduces operational fragmentation. One ticket. One path. One controlled flow of people.

Order replaces improvisation.

Stability also comes from diversification, but not the kind driven by expansion into unrelated industries. Jungfraubahn balances three seasonal pillars: the high-altitude Jungfraujoch attraction, winter sports operations, and summer “experience mountain” tourism. When one weakens, another strengthens. Snow-dependent revenue is offset by year-round international tourism. The seasons hedge each other.

The system breathes, but it does not lurch.

A quieter source of low entropy lies beneath the tracks: energy. The company operates its own hydroelectric power plant, generating electricity to power its railways and cableways. In a business dependent on elevation and transport, energy sovereignty removes a layer of external volatility. The trains run because the water flows.

Centralization further tightens the structure. Through a shared management entity, marketing, IT, procurement, and technical expertise are pooled across the group. What could have been a collection of isolated mountain railways becomes a coordinated organism. Standards are uniform. Decisions are aligned.

Entropy often enters when subsidiaries drift apart. Here, they move in formation.

The most powerful stabilizer, however, remains geography. The railway climbs into a UNESCO World Heritage landscape that cannot be replicated or relocated. There is no substitute mountain next door. No alternative “Top of Europe” to be built by a competitor. The asset is not scalable—but it is irreplaceable.

That distinction is subtle but important.

Jungfraubahn does not grow by multiplying locations. It grows by improving the experience of reaching the same place. Capacity projects like the V-Cableway increase flow and reduce waiting times, but the destination remains unchanged. The company evolves around a constant core.

In a world where businesses chase flexibility, Jungfraubahn is anchored to stone.

Some companies endure because they control habits.

Others endure because they control rules.

Jungfraubahn endures because it controls access to something that does not move.

When the destination is fixed, the system has fewer ways to unravel.

Conclusion

In the first part, low entropy emerged from discipline and inevitability.

Here, it emerges from boundaries, habits, and geography.

OMA operates inside fixed rules.

Diageo aligns with rituals that change slowly.

Jungfraubahn controls access to a place that does not move.

Different industries. Different risks. Different growth paths.

But the same underlying pattern.

Each business is structured around forces larger than management decisions. Contracts. Culture. Stone.

In investing, we often look for innovation, disruption, and acceleration.

Perhaps we should also look for gravity.

The forces that keep systems in place.

Because when the foundation moves slowly, the business above it often does too.

And in the long run, stability can be just as powerful as speed.

If you’d like to keep reading pieces like this one, I’d be glad to have you as a subscriber.

Ezra’s Notes is free, and it helps me write more essays worth your time.

Subscribe now

They fail because they become disordered.

Decisions pile up. Products multiply. Exceptions turn into rules. Complexity creeps in slowly, almost invisibly, until the system can no longer hold itself together. What once worked begins to require constant attention, constant adjustment, constant energy.

In physics, there is a word for this tendency: entropy. Left unattended, systems drift toward disorder. Business is no exception.

Yet some companies seem oddly resistant to this pull. They do not need to reinvent themselves every few years. They do not chase every new trend. They do not depend on precise forecasts or heroic execution. Their operations feel calm. Repetitive. Almost boring.

And because of that, they last.

These are low-entropy businesses. Not because they avoid change altogether, but because their core systems are built around forces that change very slowly—or not at all. Human rituals. Biological cycles. Structural simplicity. Fixed rules.

This essay is not about valuation, growth rates, or competitive positioning in the traditional sense. It is about structure. About why certain businesses remain ordered while others gradually unravel, even when they appear successful on the surface.

To explore this idea, we will look at three very different companies:

1. A funeral services provider built around an unavoidable human moment.

2. A materials company that operates at the pace of trees.

3. A retailer that wins by removing decisions rather than adding them.

They share little in common at first glance. Different industries. Different geographies. Different customers. But underneath, they are governed by the same principle: they are designed for a world that doesn’t need to change very much.

Service Corporation International (SCI) — A business built for a world that doesn’t change

The phone rings.

Not because the economy is booming.

Not because rates moved.

Not because a new technology launched.

It rings because someone just lost a parent.

That is where SCI begins. And that single fact explains almost everything about why it is a low-entropy business.

Most companies wake up every morning fighting uncertainty. Will demand hold? Will customers switch? Will a new product make them irrelevant? SCI doesn’t play that game. Its core service is triggered by an event that ignores markets, innovation cycles, and quarterly narratives. Death does not wait for better conditions.

That inevitability is the first source of order.

But inevitability alone does not create a durable business. What matters is what happens after the phone rings. And here is where the structure becomes revealing.

At the moment families walk through the door, they are not looking for novelty. They are looking for familiarity. Ritual. Reassurance. The service delivered today looks almost identical to the service delivered decades ago. Same sequence. Same steps. Same expectations. Human emotions change very little at moments like these, and SCI has built its entire operation around that constancy.

Repetition is not a weakness here. It is the system.

In most industries, repetition signals stagnation. In this one, it creates calm. Processes become muscle memory. Decisions shrink. Variation disappears. The business does not need to experiment; it needs to execute. Thousands of times. In thousands of locations. With very little improvisation.

That repetition drains entropy from the system.

There is another detail that rarely gets attention, but it matters. Part of this business is organized before it is needed. Families plan ahead. Arrangements are made in quiet rooms, years in advance, often over coffee rather than grief. From an operational perspective, this shifts time itself. Work is done today for an event that will happen tomorrow. Uncertainty gets pulled forward and neutralized.

Time, which usually destabilizes businesses, becomes an ally.

Compare that to industries obsessed with speed. Software companies sprint to outrun the next update. Media businesses chase attention that evaporates overnight. Even retailers must constantly adjust to shifting tastes. SCI faces none of that pressure. There is no technological shortcut around a funeral. No app can replace presence. No platform can compress the moment.

Change arrives slowly here. Almost politely.

This is what low entropy looks like in practice. Not growth charts or scale metrics, but a system where the number of decisions that truly matter is small—and shrinking. Where the core service fulfills a fundamental human need that remains constant across generations. Where the business survives not by anticipating the future, but by being largely indifferent to it.

There is something almost uncomfortable about that stability. In a culture that celebrates disruption, SCI feels anachronistic. Quiet. Unambitious. And that is precisely its strength.

Some businesses are designed for a world that accelerates.

Others are designed for a world that doesn’t.

Service Corporation International belongs firmly to the second category. Its advantage is not innovation, brand, or even scale. Its advantage is that the rules of its game barely move. And when the rules don’t move, order compounds.

Near the end of his life, Charlie Munger once remarked how little time he had left, not how much he had accumulated. That instinct—to see the world as finite, not scalable—is rare in business.

SCI operates as if it understands that instinct.

And in doing so, it has built something surprisingly durable: a business that wins not by changing faster, but by needing to change less.

Corticeira Amorim — A business that moves at the speed of trees

The tree is not cut.

That detail matters more than it sounds.

Every nine years, a worker peels the bark from a cork oak using a small axe. The tree stays standing. It keeps growing. It keeps absorbing carbon. And it will be harvested again. And again. For more than a century.

That single biological constraint sets the tempo for Corticeira Amorim. And it explains why this is one of the lowest-entropy businesses you will ever encounter.

In a world obsessed with speed, cork refuses to hurry.

Corticeira Amorim is the global leader in cork processing, but that description misses the point. This is not a manufacturing story. It is a time story. The raw material grows slowly, regenerates slowly, and improves with age. No software update can accelerate it. No breakthrough can compress it. The supply chain moves at the pace of biology.

That pace imposes order.

The company’s structure mirrors this constraint. Corticeria controls the entire lifecycle of cork, from forest management to final applications. Forests are managed decades ahead of demand. Raw material is prepared centrally. What can’t become a wine stopper becomes flooring. What can’t become flooring becomes insulation. What can’t become insulation becomes energy to power the factory itself.

Nothing is rushed. Nothing is wasted.

This is where most businesses introduce entropy: leftovers, exceptions, inefficiencies. Amorim does the opposite. Every fragment has a destination. Even cork dust is burned as biomass to fuel production. Waste does not accumulate because the system does not allow it to exist.

Order is engineered into the process.

At the surface, the company looks diversified. Wine stoppers. Construction materials. Aerospace components. Transportation. Energy. Sports surfaces. Fashion. Toys. But this is not diversification in the financial sense. It is reuse at scale. One material. One biological input. Many outcomes.

Think of cork as nature’s high-tech foam. Lightweight. Compressible. Insulating. Fire resistant. And endlessly recyclable. Corticeira doesn’t chase new markets; markets keep finding new uses for cork. From sealing a bottle of wine to protecting a spacecraft during re-entry, the material stays the same. Only the application changes.

That stability matters.

Innovation exists here, but it plays a different role. R&D is not about reinventing the business. It is about refining the material. Cleaning it better. Shaping it differently. Adapting it to new technical requirements without altering the underlying system. Innovation reduces entropy instead of introducing it.

The company’s sustainability narrative often gets framed as ESG. That misses the deeper point. Sustainability is not an overlay. It’s a consequence. When your raw material regenerates itself, and your process extracts value from every molecule, environmental alignment stops being a tradeoff and becomes a structural advantage.

The forests outlast management teams. The trees outlive product cycles.

This is why Corticeira Amorim feels almost immune to the forces that destabilize most businesses. There is no sudden obsolescence risk. No platform shift. No winner-takes-all dynamic. The competitive advantage is not technological or financial. It’s temporal.

The business changes slowly because it has to. And because it has to, it stays ordered.

Most companies fight entropy by adding layers: systems, dashboards, committees. Corticeira avoids entropy by accepting a simple truth: some resources cannot be accelerated. When you build around that constraint instead of fighting it, the entire organization settles into a natural rhythm.

Trees grow. Bark regenerates. Value compounds.

In investing, we often talk about businesses that scale. Corticeira Amorim is different. It endures. And in a world where everything seems designed to break faster, endurance may be the rarest advantage of all.

Costco — A business that wins by saying no

The store feels unfinished.

The concrete floor is bare.

Boxes sit on pallets.

Signs look temporary.

There are no aisles designed to inspire browsing.

And yet, every cart is full.

That contradiction is the first clue to understanding why Costco is a low-entropy business.

Most retailers try to win by offering more: more choice, more brands, more promotions, more personalization. Costco does the opposite. It removes options. Relentlessly. And in doing so, it creates order.

Walk into a typical supermarket and you face tens of thousands of decisions. Which brand? Which size? Which promotion? Costco cuts that chaos down to fewer than four thousand items. Often, just one version of a product. Sometimes two. Rarely more.

The customer doesn’t decide. Costco already did.

That is the system.

The membership card makes this even clearer. You pay before you shop. The transaction begins long before you reach the checkout. That single step flips the entire economic relationship. Costco no longer needs to extract profit from every item. The annual fee covers most of the overhead. Everything else can be sold at razor-thin margins.

The incentive to confuse disappears.

This is where entropy usually enters retail: markdowns, seasonal resets, promotional noise, constantly changing layouts. Costco avoids all of it by anchoring the model around trust. Members don’t come to discover new things. They come to buy known things at fair prices, again and again.

Repetition becomes a feature, not a flaw.

Behind the scenes, the same philosophy applies. Warehouses are simple by design. Products stay on the pallets they arrived on. Inventory moves fast. Handling is minimized. Complexity is treated as a cost, not a competitive advantage. The fewer steps something takes, the fewer things can go wrong.

Even cash behaves differently here. Inventory often sells before suppliers are paid. Time works in Costco’s favor, quietly smoothing the operation. Nothing flashy. Nothing heroic. Just a system that compounds small efficiencies day after day.

The same logic extends to people. Employees stay. Careers last decades. Managers are promoted from the floor. Knowledge accumulates instead of leaking out. In an industry known for churn, stability becomes another form of cost control.

What makes Costco unusual is not scale or pricing power. Plenty of companies have those. What makes it unusual is discipline. The discipline to say no to selection. No to aesthetics. No to short-term margin expansion. No to optimization that adds friction.

Most businesses accumulate entropy by adding layers. Costco reduces it by subtracting them.

The result is a company that barely needs to change. The core rules have held for decades. Members trust the price. Suppliers respect the volume. Employees understand the culture. The system runs because it is simple enough to be understood and repeated.

That simplicity is hard to copy. Not because it is complex, but because it requires restraint.

Where retailers fight for attention, Costco operates like a cooperative with a building. It doesn’t try to excite you. It tries to serve you. And because it does, the business remains calm while everything around it accelerates.

Some companies grow by expanding choice.

Others endure by removing it.

Costco belongs to the second group. Its low entropy comes not from innovation, but from consistency. From a refusal to complicate what already works. From the quiet confidence that fewer decisions—made well—are better than many decisions made often.

Conclusion

We often talk about business success as if it were a race against time. Faster growth. Faster innovation. Faster response. The assumption is that endurance comes from speed.

The companies in this essay suggest something different.

They endure because they are built around forces that do not accelerate easily. Death does not move faster. Trees do not grow on command. Habits formed through trust do not need constant redesign. These businesses do not fight time; they align with it.

Their advantage is not superior forecasting or clever strategy. It is restraint. The discipline to accept limits. The patience to repeat what works. The confidence to leave certain things untouched.

Low entropy is not about avoiding change. It is about deciding where change is allowed to happen—and where it is not.

Most organizations accumulate disorder by adding. New products. New processes. New exceptions. Over time, the system demands more energy just to stay in place. These businesses do the opposite. They remove decisions. They narrow variation. They build structures that can be run without constant intervention.

That is why they feel calm.

In investing, we spend enormous effort trying to predict what will change next. Perhaps we should spend more time identifying what is unlikely to change at all. Human rituals. Biological cycles. Simple rules that scale quietly.

Durability is not always about doing more.

Sometimes, it is about needing less.

And in the long run, order compounds just as reliably as growth.

If you’d like to keep reading pieces like this one, I’d be glad to have you as a subscriber.

Ezra’s Notes is free, and it helps me write more essays worth your time.

Subscribe now

Todd Combs is leaving Berkshire Hathaway and GEICO to join JPMorgan.

A quiet, almost monastic investor stepping away from the house Buffett built— that alone is enough to make any long-term thinker pause.

According to CNBC, JPMorgan hired him to lead its new Security and Resiliency Initiative, looking for direct equity investments across defense, aerospace, healthcare, and energy.

Yes, the industries that matter when the world becomes unpredictable.

When someone like Todd moves, he isn’t chasing headlines.

He’s redirecting a process.

A worldview.

A philosophy.

And that philosophy might be one of the clearest frameworks for thinking about businesses and investing today.

Below are the timeless lessons I’ve taken from his writing — particularly from his introduction to Security Analysis — and from studying how he actually works. Because if you understand Todd’s process, you understand why he succeeds in environments where others get lost in noise.

Start With Facts, Never With Stories

The first principle of Todd’s philosophy is almost austere:

When done well, investing involves learning how to process information in order to determine when the odds are in your favor; the goal is to make educated bets based on facts and not stories.

That sentence alone explains why Buffett hired him.

Todd avoids the trap that captures so many smart analysts: beginning with a narrative and then gathering evidence that confirms it.

He works in the opposite direction.

His process always starts with primary sources:

• SEC filings

• Annual reports

• 10-Ks, 10-Qs

• Trade journals

Never with:

• Management narratives

• Sell-side research

• “Channel checks” from biased insiders

• Consensus stories

Opinion last. Facts first.

Because once you adopt a story, you unconsciously begin protecting it.

Great Businesses Reveal Themselves Through Their Economics, Not Their Story

Todd breaks analysis into the first of his three major buckets:

1. Find a good business

What does “good” actually mean?

According to Todd, a great business has:

• A wide moat

• Low capital intensity

• Pricing power

• Recurring revenues

• Staying power

• The likelihood of long-term growth

And you can only confirm these traits by ripping the business “down to its studs.”

His sequence is always quantitative, then qualitative:

a) Start with the balance sheet

The balance sheet tells you how the business funds itself, how much risk it carries, and how fragile it truly is.

b) Cash flow statement second

Cash flow reveals how much oxygen the business produces without illusions.

c) Income statement third

Think in terms of a dollar of revenue that flows into a business and runs through the cash flow statement, then the balance sheet, and last the income statement. This is ultimately how we derive a company’s return on invested capital, which is usually a reliable shorthand for the quality of the business.

Todd doesn’t want to know how a company performed.

He wants to understand how each unit performs — a store, a subscriber, a policy, a widget.

It’s business reality, not spreadsheet reality.

Start with facts and not opinions. If you start with opinions, it’s very easy to become wedded to them even when the facts run counter to the popular narrative.

The Second Bucket: Great Management Actually Matters

Most investors underestimate management. Todd does not.

The importance of good management is almost universally underestimated, yet it is one of the most crucial determinants of a company’s intrinsic value. As Graham and Dodd said, “You cannot make a quantitative deduction to allow for an unscrupulous management; the only way to deal with such situations is to avoid them.”

He evaluates leadership as if he were buying the entire company himself.

How does Todd judge management?

Not by charisma.

Not by interviews.

Not by investor days.

He studies:

• Incentives (via the proxy statement)

• Capital allocation decisions

• Use of leverage

• Whether the CEO acts like an owner

• Whether they plant seeds for the future or borrow from it

He goes deeper through scuttlebutt — talking to people who have actually worked with or for the CEO.

• Are they honest?

• Do they break teams?

• Do they oversimplify?

• Do they confront reality or avoid it?

In Todd’s eyes, CEO incentives predict outcomes better than CEO words.

While we know that a long-term perspective is crucial in running a business, there are often unfortunate incentives to become overly short-term oriented. The private owner of a great business isn’t worried about quarterly earnings, meeting market expectations, stuffing sales channels, pursuing aggressive accounting treatments, or withholding long-term investments to improve reported short-term results.

The Third Bucket: Price Still Matters (More Than Ever)

A great business with a great CEO is still a bad investment at the wrong price.

This is where Todd’s discipline shows.

He insists on three valuation ideas:

1. Owner earnings > EBITDA: EBITDA, especially adjusted EBITDA, is a shortcut Todd distrusts. He focuses on owner earnings — the cash left after maintaining the business.

2. Capital structure matters. A company dependent on leverage is fragile, even if earnings look strong.

3. Margin of safety. Todd reminds us: We know far less than we think.

Perfect information does not exist; there are only confidence intervals. This is, of course, at the heart of why a margin of safety matters so greatly in investing. If you start with the premise that there is only so much one can know, of course you need a margin for error. The less you know, the greater the margin needed.

What We Can Learn From Todd Today

Here are the lessons I return to most often:

1. Start with facts. End with opinions. Never reverse the order. The moment you start with a story, you’ve already lost.

2. A spreadsheet is not business reality. Cash flow, incentives, and unit economics are reality.

3. Management quality is not a narrative — it’s a set of behaviors. Incentives > charisma > presentations.

4. The right price matters more than the right narrative. Good businesses are terrible investments at the wrong prices.

5. Humility compounds. You cannot know everything. But you can design a process that works despite that.

Conclusion

I believe Todd Combs didn’t leave Berkshire to become a different investor.

He left to express the same principles in a world where those principles are desperately needed.

And that’s the point:

A great investor doesn’t chase outcomes.

A great investor builds decisions.

Decisions that survive uncertainty, noise, and time.

Investors must find a way to navigate through uncertainty and randomness. There are decisions beingmade every day, and sometimes made decades earlier, that are still affecting outcomes today in every business and industry.

That’s the Todd Combs philosophy.

And there has never been a better moment to study it.

I’d love your take: what’s the most important idea you’re taking from this essay?

Let me know in the comments.

If you’d like to keep reading pieces like this one, I’d be glad to have you as a subscriber.

Ezra’s Notes is free, and it helps me write more essays worth your time.

Subscribe now

Glad to have you here!

Ezra.

Stealing an idea from my friend this edition is all about visuals.

And in a way, it works as a companion piece to my recent essay on the AI mania—where I argued that every technological boom is really a story about how humans react to accelerating change.

If that essay explored the narrative side of the cycle, this post looks at the evidence behind it.

The scale of capex, the shifting economics of hyperscalers, the power constraints, the geographic and sector winners, the capital loops between OpenAI, Microsoft, Nvidia and everyone orbiting them… When you see the data laid out, the magnitude of the moment becomes much clearer.

So today’s post is simple: 19 visuals that explain the AI boom better than a thousand takes.

Capex, returns, sectors, beneficiaries, risks—everything in one place.

Let’s dive in.

If this added value to your day, give it a like and share it with someone who enjoys long-term thinking. Thanks for supporting the newsletter.

how we react when the future seems to accelerate faster than our ability to understand it.

AI triggers hope, fear, ambition, and vertigo in equal measure. And from the inside, everything feels inevitable, logical, even obvious… until it no longer is.

In recent weeks, I read four Q3 2025 letters written by investors I respect —Wedgewood Partners, Praetorian Capital, Greenlight Capital, and Hayden Capital— each of them observing the same fever from a different angle.

Rather than trying to decide who was “right,” I was interested in understanding what we could learn by comparing their perspectives.

These four letters didn’t just teach me about the current state of AI.

They taught me about the way investors interpret a mania.

Lesson #1 — (Wedgewood Partners): The Scale of the Boom

“Manias feel inevitable when the numbers get so big they stop having meaning.”

Wedgewood Partners was the first to remind me of something uncomfortable:

Behind every boom, there is a point where enthusiasm stops acknowledging arithmetic.

Their letter includes an observation that perfectly captures the current moment:

“Artificial Intelligence (AI) stocks long continue to be darlings of Wall Street… yet, to try to grasp the ever-growing forecasts of trillions in the necessary infrastructure to support the current insatiable demand does give an investor pause — particularly given how crowded the ‘AI trade’ has become.”

What’s interesting is the idea beneath that sentence:

The notion that AI is not only absorbing capital, but also the collective imagination.

To articulate that feeling, the Wedgewood team turns to an unexpected cinematic metaphor: The Blob.

Just as that creature devoured everything in its path, today’s investment in AI is creating a growing mass of CAPEX, credit, and hype—within which economic logic begins to dissolve.

The numbers explain why:

• The tech sector—traditionally “asset-light”—is mutating into a capex-heavy, debt-dependent model.

• For two decades, data centers were funded with internal cash flow; yet in 2025 alone, the tech giants raised US$150 billion in debt to sustain the AI fever.

• According to Bain & Co., the US$500 billion in required CAPEX will demand US$2 trillion in revenue to be justified.

Figures that, in Wedgewood’s words, defy analytical comprehension… and yet, the market is cheering.

But the most unsettling part isn’t the magnitude of the spending—it’s the speed.

The GPUs powering the revolution depreciate in 3–5 years, while NVIDIA is already accelerating toward annual upgrade cycles.

This means massive amounts of capital could become obsolete long before recouping their investment.

Wedgewood calls this a potential wave of depreciation-driven earnings compression.

And then there’s another constraint people discuss far less: energy.

Some data centers consume as much electricity as an entire U.S. city.

Wedgewood warns that the real bottleneck won’t be technological, but political.

Without a massive expansion of the power grid—and possibly water access—AI will run straight into the physical limits of the system.

Lesson #2 — (Harris Kupperman of Praetorian Capital): When the Math Doesn’t Add Up

How much revenue is needed to justify the current level of capex spend and give AI investors a return on their capital??

Harris “Kuppy” Kupperman of Praetorian Capital describes something unsettling.

After speaking with insiders in the data-center ecosystem, he discovered that many of them—the very people building the AI infrastructure—are confused.

They don’t understand how these projects are supposed to generate economic returns.

And the most concerning part isn’t that they don’t understand it. It’s that investors don’t either.

Kuppy agrees with Wedgewood on a critical point: GPUs become obsolete in 3 to 5 years, precisely as NVIDIA accelerates its upgrade cadence.

This means that to reach a reasonable breakeven, the industry would need roughly US$1 trillion in revenue—an amount that turns any financial model into an act of faith.

Then comes the uncomfortable question he raises: hyperscalers have spent years subsidizing money-losing businesses to pull users into their ecosystem.

Kuppy wonders whether AI is simply another one of those businesses that, strategically, can lose money indefinitely.

Part of the market believes these costs will be offset once AI replaces highly skilled jobs.

Harris isn’t buying it: that disruption—if it arrives—is years away and will require technology far more advanced than what exists today.

His diagnosis is broader: without U.S. government intervention, and without the investment boost created by AI, the United States would already be in a recession. In fact, he argues that the AI bubble may represent more than 100% of recent economic growth.

To illustrate the phenomenon, he turns to the 19th-century railway bubble: a transformative industry, enormous promise, insiders convinced that investment had to continue endlessly…

Yet the cycle still blew up multiple times, triggering financial panics and sending many companies into bankruptcy.

In his sharpest line, he compares the two eras:

Raising capital for a money-losing business is hard. Back then, you got a free train trip. Now you get the free use of an LLM.

This investor is also skeptical of the circular investment patterns among Big Tech: companies investing in one another to justify growth—alarmingly similar to what happened with fiber-optic networks during the dot-com bubble.

Finally, he warns that this bubble won’t burst because of a technical signal, an obvious metric, or a sudden narrative shift.

His conclusion is human, almost psychological:

Like all bubbles, it implodes when people get tired of funding it.

Lesson #3 — (David Einhorn of Greenlight Capital): The AI Spending Delusion

There is a good chance that 25 years from now, AI will turn out to be even more important than we currently imagine. But, the path from here to there is likely to be very bumpy for investors.

David Einhorn of Greenlight Capital arrives at conclusions eerily similar to Kupperman’s: He too sees a deep disconnect between the enormous sums being allocated to AI and the basic arithmetic required to justify them.

The most revealing part is that not even the protagonists themselves know how they’re supposed to earn back what they’re spending: Tim Cook (Apple), Mark Zuckerberg (Meta), and Sam Altman (OpenAI) have all failed to articulate a clear path to economic returns in the face of this massive wave of CAPEX.

Their answer is always the same: “This will eventually pay off.”

But the key word is “eventually,” and Einhorn wonders how long that eventually can last.

In his letter, he cites a McKinsey study projecting something monumental: by 2030, global AI-related CAPEX could reach US$6.7 trillion—with US$5.2 trillion going to data centers alone.

The question he raises is simple and brutal:

Where will those nearly US$7 trillion come from?

His analysis is straightforward: the math doesn’t add up.

Not even the “Mag Seven” —the most profitable companies on the planet— generate enough free cash flow to finance this vision of the future.

Collectively, they produce around US$500 billion in FCF per year.

Half of that is already committed to stock buybacks and dividends.

Their total book equity is roughly US$1 trillion — a fraction of the spending required.

Even if Wall Street, private equity, and venture capital emptied their coffers, there would still be a multi-trillion-dollar gap that could only be filled with massive new debt.

To justify all this spending, AI would need to generate about US$2 trillion in annual revenue by 2030—equivalent to the combined size of today’s global market for advertising and software subscriptions.

Einhorn’s question is obvious: Is that even remotely plausible?

He then dives into something even more troubling: the illusion of revenue.

He describes how much of the “AI revenue boom” doesn’t come from real customers, but from AI companies buying products from one another.

His line is precise:

Much of the revenue in AI simply comes from AI companies buying products and services from each other.

He even walks through a concrete example:

A single dollar spent on a ChatGPT subscription can turn into eight dollars of “reported revenue” along the chain—OpenAI to Microsoft, to CoreWeave, to NVIDIA—without a single dollar of net economic value being created.

It is, quite literally, the same dollar inflating multiple income statements.

This dynamic, he warns, creates a false illusion of growth while masking value destruction, overcapacity, and a fragile structure dependent on cheap capital.

Einhorn’s other critical point is the almost religious faith in AGI (Artificial General Intelligence).

He challenges the belief that we are close to machines that “think” or improve themselves.

Current models, he argues, are statistical systems that mimic patterns; they do not understand, reason, or make real inferences.

To illustrate this excess of optimism, he points to the case of self-driving cars: for a decade, they were always said to be “one year away.” And yet, that year never arrived.

The expectation of imminent AGI, he argues, is not technology.

It is wishful thinking, amplified by financial incentives.

Lesson #4 — (Fred Liu of Hayden Capital): Bubbles as Engines of Progress

Not all bubbles destroy. Some prepare the ground.

In contrast to the skeptical tone of the other managers, Fred Liu of takes a different stance:

Stop looking at the trees and step back to see the forest.

He doesn’t focus on whether we are or aren’t in a bubble.

He proposes something more interesting: to ask what kind of world is being built by this wave of AI investment.

His thesis runs counter to market intuition:

Bubbles are a feature, not a bug, of technological progress.

Instead of seeing them solely as financial tragedies, Liu sees bubbles as mechanisms of civilizational advancement: periods of excess that finance critical infrastructure, massive experimentation, and new mental models.

To explain this, he turns to three illuminating historical examples.

1. “Canal Mania”: The Infrastructure That Survived the Speculation

In the late 18th century, the British financed thousands of miles of canals in a speculative frenzy. Many companies went bankrupt. Many investors lost everything.

But the country was left with a network of 4,000 miles of canals that drastically reduced transportation costs, enabled the movement of fragile and heavy goods, and became the logistical backbone of the Industrial Revolution.

Liu highlights something fascinating:

Canal transport became cheaper after the bubble burst because the collapse erased the original capital costs and forced competition among overbuilt routes. New owners could price at marginal cost instead of the much higher, original construction cost, turning canals into near-free public goods.

The capital was lost.

The infrastructure remained.

And with it, a new economy became possible.

2. The Railroad Boom: The Excess That Unified a Country

In the late 19th century, the United States experienced another wave of euphoria: railroads.

In 40 years, the nation went from 30,000 to 200,000 miles of track—the largest railway system in the world.

It was a period of extreme overinvestment: fares fell 80%, competition was brutal, and a quarter of all rail lines went bankrupt.

But the country emerged more unified, more industrialized, and with new industries that would have been impossible without that excess:

• national catalogues (the original Amazons),

• industrial-scale agriculture,

• unified markets,

• mobility and economic expansion.

Financial chaos gave way to an unprecedented productive leap.

3. The Dot-Com Bubble: Overbuilding That Created the Modern Internet

This one is closest to our memory.

The dot-com bubble produced:

• excess fiber-optic cable,

• unused data centers,

• massive bankruptcies,

• delusional expectations,

• “dark fiber” sold for ridiculous prices.

But that excess became the backbone of the modern Internet:

• cheap broadband,

• ecommerce,

• streaming,

• SaaS,

• cloud computing,

• smartphones.

Liu describes it as a process of parallel experimentation:

New browsers, search engines, ecommerce models, and most importantly:

An entire population coming online, creating the cognitive foundation necessary for the digital economy that would dominate the following decades.

Conclusion

What changes is the technology.

What doesn’t change is how we react to it.

AI can be a bubble, an opportunity, or a revolution. It’s likely a bit of all three.

What is certain is that the challenge remains the same as always:

To think calmly in a world that keeps accelerating

I’m curious: which of the four perspectives resonated most with you?

Let me know in the comments.

If you’d like to keep reading pieces like this one, I’d be glad to have you as a subscriber.

Ezra’s Notes is free, and it helps me write more essays worth your time.

Subscribe now

Everything is condensed into clear notes you can read in minutes.

1. Industrial

• Softening fundamentals: Demand cooling and excess supply from the last boom.

• Temporary NOI boost: Rent catch-up effects from COVID surge are masking slowing growth.

• Market inefficiency: Plymouth Industrial’s buyout offer (~$24/share) exposes valuation gaps in REITs.

2. Office

• Uneven recovery: “Flight to quality” driving gains for top-tier properties.

• Leasing rebound: Sunbelt Class A assets and NYC leading recovery; AI demand lifting West Coast markets.

• Oversupply risk: Costly new developments in Midtown Manhattan could hurt landlords.

3. Multifamily / SFR

• Sunbelt weakness, coastal strength: Rent growth still negative YoY in the South but improving.

• SF rebound: Bay Area rents rising sharply amid AI-driven hiring and capital inflows.

• Housing slowdown: Single-family home starts expected to weaken further in 2H 2025.

4. Retail

• Top performer: Retail showing strongest fundamentals among CRE sectors with ~3–4% NOI growth.

• Low new supply: Minimal construction supports landlords.

• Tariff risk easing: Legal challenges have tempered the threat of new import tariffs.

Credit to

of Warden Capital for this amazing update on the CRE market.

Subscribed to his newsletter!

See you next week with more timeless notes.

Until then, keep compounding knowledge.

— Ezra

Enjoyed this?

Pass this note along to a fellow investor

Share

Everything is condensed into clear notes you can read in minutes.

These are my notes from a recent interview that David did with Chris Davis.

Enjoy!

I think all investing is value investing. And if it isn't, it's it's either gambling or speculating. And investors should recognize that the biggest threat to their generating a return over time is their own behavior.

• All true investing is value investing; anything else is speculation or gambling.

• Focus on earnings power and intrinsic business value, not market psychology.

• Growth is part of value—profitable growth adds intrinsic worth. The false dichotomy of “growth vs. value” misleads investors.

• Rejects momentum investing, which relies on rising prices rather than fundamentals.

• Let value compound gradually rather than chasing immediate payoffs.

• The greatest threat to investor returns is **their own behavior** (buying high, selling low).

• On Mag 7: Each company is unique, but high valuations require heroic assumptions (e.g., 20% growth for a decade, >50% margins).

• Don’t predict, prepare — invest with the expectation of enduring shocks.

• Not just buying “$1 for 50¢” but preferring durable, compounding businesses even at fair prices.

• A growing, resilient company may justify paying more upfront.

• Avoid emotional trading, or hire an advisor who enforces patience.

• Systematic investing (regular contributions over time) is more powerful than trying to time markets.

• Value investing suits those with patience, discipline, and a research mindset.

• Success comes not from being right often, but from making outsized gains when right and limiting losses when wrong.

I had a business card, it would say analyst on it that that's still my full time job. It's a part of the job I love. I love visiting companies. I love, you know, this idea of business people. I spend a lot of time visiting companies. I love that that part part of the job. I like going out and traveling.

See you next week with more timeless notes.

Until then, keep compounding knowledge.

— Ezra

Enjoyed this?

Pass this note along to a fellow investor

Share

Everything is condensed into clear notes you can read in minutes.

These are my notes from Howard Mark’s latest memo.

Enjoy!

Howard Marks - The Calculus of Value

• Value is derived from fundamentals and earning power (current and future).

• Price is what investors pay, largely determined by psychology, optimism vs. pessimism, and sentiment.

• Investment success depends on paying the right price relative to true value.

• Short-term market moves are driven more by sentiment than fundamentals.

• Price tends to gravitate toward value in the long term, but can diverge wildly in the short term.

• Bubbles and crashes are born when psychology pushes prices to extremes.

• S&P 500 forward P/E near 23, well above historical averages.

• Historical data suggests such valuations often result in weak (~0–2% annually) long-term returns.

• Market cap-to-GDP ratio (Buffett Indicator) at record highs.

• Stocks trading at 3.3x sales, another all-time high.

• Signs of euphoria: high valuations, narrowing credit spreads, and “meme stock” activity (again).

• Risk tolerance elevated; spreads on corporate debt near historic lows.

• Relief rallies fueled by optimism despite tariffs, inflation risk, and debt concerns

• Strong recent gains concentrated in the “Mag 7” tech stocks (Apple, Microsoft, Alphabet, Amazon, Meta, Nvidia, Tesla).

• AI enthusiasm and belief in U.S. market supremacy are reinforcing bullish sentiment.

• Optimism and FOMO outweigh concerns about tariffs, inflation, and deficits.

• Fundamentals have deteriorated (tariffs, inflation, U.S. fiscal concerns), yet asset prices are higher.

Compared to the past, today’s S&P 500 is increasingly made up of companies that (a) grow faster, (b) are less cyclical, (c) require less incremental capital to grow, enabling them to generate more free cash flow, and (d) have much stronger competitive positions or “moats.” Thus, they deserve above average p/e ratios.

Read the full memo

See you next week with more timeless notes.

Until then, keep compounding knowledge.

— Ezra

Enjoyed this?

Pass this note along to a fellow investor

Share

Everything is condensed into clear notes you can read in minutes.

These are my notes from Todd Comb’s “Introduction to Part IV” of the new edition of Security Analysis.

Enjoy!

In investing, and in life for that matter, you can’t choose outcomes, but you can choose the decisions that may get you the outcomes you want.
— Todd Combs

• The goal in investing is to make educated bets based on facts and not stories.

• A great analyst has the ability to keep things simple and determine what matters most. Its willing to rip a company down to its studs and understand each part before they reassemble it.

• A spreadsheet is not business reality.

Todd’s three buckets

1. Find a good (great) businesess

In determining if a business is great, I like to start on the quantitative side: I focus on the balance sheet, accounting practices, and unit economics, and then turn to cash flow generation. Meanwhile, the qualitative side involves reference calls to assess the quality of management and channel checks to reveal how a company’s products are selling in real time.
— Todd Combs

According to Todd, a good (great) business has:

• A wide moat

• Low capital intesity

• Pricing power

• Recurring revenues

• Staying power

• The likelihood of long-term growth

Investment process

• Start with facts and not opinions. “If you start with opinions, it’s very easy to become wedded to them even when the facts run counter to the popular narrative.”

• SEC filings, annual reports, trade magazine articles (NO management or colleague’s narratives or sell-side reports.

• Don’t look at earnings (worse: adjusted earnings)

• Instead, start with the balance sheet and the cash flow statement.

• Think about the business in terms of unit economics: “Think in terms of a dollar of revenue that flows into a business and runs through the cash flow statement, then the balance sheet, and last the income statement.”

• Review the last 10 years and examine the changes in retained earnings, debt, and overall capital intensity as compared to the top-line growth of the business itself.

2. Find good (great) management teams

The importance of good management is almost universally underestimated, yet it is one of the most crucial determinants of a company’s intrinsic value.

— Todd Combs

• Don't assess management based on conventional marketplace narratives. Go deeper.

• When analyzing management's track record, examine their incentives; how the spend their time, and perform scuttlebutt research. (Not presentations or meetings with management).

• Persistent tailwinds can also make a corporate management team seem smarter than they are.

Good capital allocation is crucial; a wonderful business and a talented CEO is not enough to offset poor capital allocation

• Yes, capital allocation is one of the most important parts of the track record that must be examined.

• You want management teams focused on long-term and not short-term results (not quarterly earnings, meeting market expectations, stuffing sales channels, pursuing aggresive accounting treatments, or withholding long-term investments to improve reported shor-term results).

• CEO focused on running the business, not spending months on the road promoting the company or meeting with investors.

Ask yourself, if this were your family business, wouldn’t you want your CEO single-mindedly focused on running the business? I want to back CEOs who are focused on substance, not collecting style points.

• Scuttlebutt research on CEOs: speak to current and former executives who have either reported to the CEO. The goal? Go deeply into their personalities to gain context. Are they toxic? Are they intense as a leaders?

A quick checklist to assess management's incentives

Read the proxy statements and look for changes year to year:

• Is the basis for management compensation and stock option awards sensible?

• Are they being measured for enduring accomplishments or instead for favorable short-term share price fluctuations that could be purely the result of luck or even manipulation?

• Do they reward actual returns on capital or simply profitless growth?

• Do they create asymmetric upside for the CEO for taking outsized risks?

• Is the CEO regularly selling his or her shares for “personal reasons,” or are they acting like a true owner and fiduciary?

• If the company were private, how different would the structure of the CEO’s compensation and incentives be?

3. Find a the “right” price

The best businesses are anti-fragile; that is, they continue to thrive and even deepen their moats in times of adversity and volatility, such as when competitors get in trouble and their best customers and employees may be up for grabs.

— Todd Combs

• Start by your own. Do not base your analysis on someone else's opinion, story, conference calls. "he danger of this process is that it often leads to the formation of an opinion based on someone else’s perception and analysis, rather than your own."

• Assess a moat. Not the quick and easy way (long track record of high ROIC); go deeper to see if the moats of the company are strong or porous.

• Business moats are not static and come with different characteristics—brands, low costs, convenience, and network effects all constitute real but very different moats.

• Do channel checks: talk to customers, suppliers, and former employees. Ask yourself: " if I were considering putting 100% of my net worth in this business, what would I want to know?"

Moats checklist

• Does the business have a sustainable and/or expanding moat?

• What are the weakest links over the next five years?

• Are there hidden path dependencies?

• Does the business have pricing power, and how has it been exercised?

• How (anti) fragile is the business, and will it thrive in the next downturn?

• What would it take to replicate this business?

• Will the business be in a better position with an even wider moat five years from now?

Remember that simple is usually better; in investing, there is no extra credit for degree of difficulty.

Valuation

• A business is worth the sum of its discounted cash flows in perpetuity.

• Find a capital light business that can grow for prolonged periods of time.

• The worst business is one taht grows and consumes increasing amounts of capital at returns that fail to exceed its cost of capital.

Great businesses can be terrible investments at the wrong price, and an average business can be a great investment at the right price. Fast-growing businesses can be attractive, but only at the right price.

• It's not growth at all costs. Price still matters.

• Always have a margin of safety

• Focus on owner earnings, NOT EBITDA (or even worse - Adjusted EBITDA)

• Take into account the company's capital structure.

To finish, some Todd’s great quotes

Human beings constantly process complex information about the world and simplify it to make sense of it and put it into context.

Investors attempt to look behind the curtain of a company’s operations, but they often realize how little they actually know.

Even extensive research will uncover only a small fraction of what one can possibly know about a business.

Investors must find a way to navigate through uncertainty and randomness. There are decisions being made every day, and sometimes made decades earlier, that are still affecting outcomes today in every business and industry.

Perfect information does not exist; there are only confidence intervals. This is, of course, at the heart of why a margin of safety matters so greatly in investing. If you start with the premise that there is only so much one can know, of course you need a margin for error. The less you know, the greater the margin needed.

See you next week with more timeless notes.

Until then, keep compounding knowledge.

— Ezra

Enjoyed this?

Pass this note along to a fellow investor

Share