The one company we could never analyze — because we quote its chairman in every letter
Every analysis on this site is written in a borrowed voice. When we weigh a moat, invoke a circle of competence, ask whether a business is wonderful or merely fair, warn against overpaying for growth or reaching for a dividend a company cannot afford — we are channelling one man, Warren Buffett, and his late partner Charlie Munger. It has always felt slightly improper, then, that we had never turned the lens on the master's own company. Partly reverence. Mostly a technical problem: how do you value a business that is itself a machine for valuing businesses? Berkshire is not a company so much as a portrait of a mind — sixty years of Buffett's capital-allocation decisions, fossilised into a $1.08 trillion holding company. To analyze Berkshire is to analyze the analyst.
We finally do it now for one reason: 2026 is the first year in sixty that Berkshire is not being run by Warren Buffett. On 1 January 2026, Greg Abel became Chief Executive; on 28 February 2026, Abel — not Buffett — signed the annual shareholder letter, the first not written by Warren since 1965. And in his first quarter with his hands on the equity portfolio, Abel did something that would have been unthinkable under the old regime: he sold Visa, Mastercard, Amazon and UnitedHealth, and tripled Alphabet. The company that canonised "our favorite holding period is forever" suddenly traded with a velocity its founder never showed. Fortune's verdict: "It's not Buffett's Berkshire anymore." Whether that is a tragedy or a graduation is the single most important question in the market, and the whole of this report.
"Our favorite holding period is forever." — Warren Buffett, 1988 letter — the sentence Abel's first year appeared to contradict
Strip away the mystique and what is Berkshire? Three things bolted together. First, a colossal insurance operation — GEICO, General Re, Ajit Jain's reinsurance — that collects premiums today and pays claims years later, and in the meantime gets to invest the ~$176 billion of "float" in between, generally getting paid to hold it. Second, a collection of wholly-owned real businesses — the BNSF railroad, Berkshire Hathaway Energy, and a manufacturing-and-retail empire from See's Candies to Precision Castparts. Third, a $350-billion portfolio of stocks — Apple, American Express, Coca-Cola, Bank of America. The float funds the businesses and the stocks; the businesses and stocks throw off cash; the cash funds more of everything. It is the most elegant compounding machine ever built, and for sixty years one man decided where every dollar went. Now someone else does.
| Founded | Textile mill (1839); Buffett took control 1965 · HQ Omaha, Nebraska |
| Sector / Industry | Financial Services · Diversified insurance holding company (a conglomerate) |
| CEO | Greg Abel (since 1 Jan 2026) · Warren Buffett remains Chairman · Ajit Jain, VC Insurance |
| Makes money from | Insurance float + wholly-owned businesses (rail, energy, manufacturing) + a ~$350B stock portfolio |
| Revenue (FY2025) | ~$371B · operating earnings ~$44.5B · net income swings on mark-to-market equities |
| Market capitalisation | ~$1.08 trillion · ~$397B in cash · pays no dividend (by design) |
How a failing textile mill became the greatest compounding machine on Earth
Berkshire's origin is the most instructive mistake in investing history, and Buffett tells it against himself. In the early 1960s he was buying shares of a dying New England textile mill — a classic Graham "cigar butt," bought cheap for "one last puff." In May 1964 the mill's owner, Seabury Stanton, verbally agreed to buy Buffett's stake back at $11.50, then sent the tender at $11.375 — "he chiseled me for an eighth," Buffett fumed. So out of spite Buffett bought more, took control of the whole doomed enterprise, and fired Stanton. "I had become the dog who caught the car," he wrote fifty years later. He then spent two decades watching the textile operation bleed — it finally shut in 1985, its looms auctioned for $26 apiece as scrap — while he quietly redirected its cash into the thing that actually mattered.
| Year | Milestone |
|---|---|
| 1964–65 | Spite purchase: Buffett takes control of a failing textile mill. He later calls it his worst investment decision — the capital would have compounded far more inside a good insurer. The textile business is a permanent monument to buying cheap-and-bad over wonderful-and-fair. |
| 1967 | The pivot: Berkshire buys National Indemnity from Jack Ringwalt for $8.6M — 'Fifteen minutes later we had a deal. No audit, no lawyers.' Insurance float becomes the engine. Also: Berkshire pays its ONLY dividend ever — 10 cents — Buffett jokes he 'must have been in the bathroom.' |
| 1972 | See's Candies ($25M) — Munger's turning point. Buffett 'gulped' at three times tangible book; Munger insisted a wonderful brand was worth it. It returned >$2B pre-tax. The lesson — 'a wonderful business at a fair price beats a fair business at a wonderful price' — became the house religion. It is the religion of this website. |
| 1988 | Coca-Cola: 'only in the summer of 1988 did my brain finally establish contact with my eyes.' A $1.3B stake becomes one of the great holdings, still owned today. The 'forever' philosophy is coined. |
| 1996–2009 | GEICO bought whole ($2.3B); General Re; then BNSF in 2009 for ~$44B — 'an all-in wager on the economic future of the United States. I love these bets.' Berkshire becomes an owner of the American economy itself. |
| 2016 | Apple — the great late-career triumph, now the largest holding. Proof Buffett could still evolve: from a man who avoided tech to one whose biggest position was a phone company he understood as a consumer brand. |
| 2023 | Charlie Munger dies at 99, 33 days short of 100. 'Berkshire could not have been built without Charlie.' The architect is gone; the general contractor carries on alone — briefly. |
| 2026 | Greg Abel becomes CEO (1 Jan); writes the annual letter (28 Feb) — the first not signed 'Warren' in 60 years. Buffett stays Chairman. A new era begins. |
The through-line matters for the valuation. Berkshire's greatness was never the textile mill, the railroad, or even the float — it was a man's judgment about where to move a dollar next, exercised with discipline over sixty years and a longer runway than any other capital allocator ever got. That judgment is the asset that is now, gradually, being replaced. Everything that follows — the moat, the balance sheet, the price — has to be read through that one transition.
Simple idea, sprawling internals
The machine, in five steps:
Berkshire sits one notch below the Coca-Colas and Visas of our shelf on understandability, and the reason is honest: it is a conglomerate, the very structure Buffett spent his career warning was usually a device to hide mediocrity. His defence was always that Berkshire was the exception — a conglomerate whose glue was not financial engineering but trust and capital-allocation genius. That defence rested on him personally. The ten-year future of GEICO, BNSF and Coca-Cola is highly knowable; the ten-year future of how well the next person allocates the firehose of cash they produce is the genuine unknown, and it is not a small one.
Float is the magic · four engines underneath
Berkshire's revenue (FY2025 segment disclosure) shows the sheer breadth — but the magic is invisible on the revenue line, because it lives in the balance sheet, as float.
Why float is the whole story. Imagine a bank that pays you to hold your money instead of charging you. That is float. Berkshire's insurers collect ~$176 billion of premiums they will eventually pay out as claims — but "eventually" can be many years, and in the meantime that $176 billion is Buffett's (now Abel's) to invest, and because Berkshire's underwriting is usually profitable, it is effectively being paid to borrow it. That is the trick behind the whole 60-year record: permanent, interest-free, growing leverage, invested by the best allocators alive. On top of the float sits the second layer of genius — the wholly-owned businesses and the stock portfolio don't just earn; their retained earnings compound inside Berkshire tax-efficiently, a snowball rolling down a very long hill. The problem in 2026 is that the hill has flattened: at $1.08 trillion, the snowball is so large that only a handful of acquisitions on Earth can move it, and the cash keeps piling up because there is nowhere big enough to put it.
A fortress of moats — plus one that only Berkshire has
Berkshire's moat is not one wall but a fortress built of many, plus a structural advantage no competitor can copy:
1 · Permanent, cost-free capital (the float). ~$176 billion of interest-free investable money that grows with the insurance book. No hedge fund, no private-equity firm, no bank has anything like it — they all rent their leverage; Berkshire is paid to own its. This is the deepest moat in finance.
2 · The "seller's brand." When a family-owned business wants to sell to a permanent home that won't strip it, fire its people, or flip it — there is exactly one buyer with that reputation, built over sixty years: Berkshire. That reputation lets it buy great businesses without auctions, at fair prices, from owners who care where their life's work lands. It is a moat made of trust, and it is uncopyable because it took six decades to earn.
3 · Fortress balance sheet as a weapon. $397 billion of cash is not just safety; it is optionality. In 2008 and 2020, when everyone else was forced to sell, Berkshire was the buyer of last resort, dictating terms (the Goldman and BofA preferreds). The balance sheet lets Berkshire be greedy precisely when others are fearful — the moat that only shows up in a crisis.
4 · The underlying businesses' own moats. BNSF (rail duopoly), BHE (regulated monopoly utilities), GEICO (low-cost insurance), See's (brand), Precision Castparts (aerospace criticality) — Berkshire is a portfolio of moats, diversified across the whole economy.
Where is the crack? Not in the walls — it is at the top. Berkshire's ultimate moat was always the judgment of the man deciding where the capital went, and that is precisely the input now changing hands. The fortress is intact; the question is who holds the keys, and whether the "seller's brand" and the discipline survive without Buffett's personal signature on them. I score the moat a 9 — one of the widest collections of advantages in existence, marked down from a 10 only by the key-man transition sitting on top of it.
The paradox at the heart of Year One
Here is the tension that makes 2026 the year to finally analyze this company. The firm that taught the world "our favorite holding period is forever," that mocked "hyperactive" trading, that held Coca-Cola for 38 years — spent Abel's first quarter as portfolio manager doing the opposite. In one 13F it exited sixteen positions, cutting the book from ~40 holdings to ~26, and the names it sold are uncanny to a reader of this site: Visa and Mastercard (the payments duopoly we rate a 10-moat), Amazon, UnitedHealth, Domino's. And it tripled Alphabet — a growth-and-AI bet from the firm most wary of technology — and, most startling of all, bought an airline (Delta), six years after Buffett swore off airlines forever having dumped all four in 2020.
| 'It's not Buffett's Berkshire' (the worry) | 'The philosophy is intact' (the reassurance) |
|---|---|
| Un-Buffett velocity — 16 exits in a quarter from the 'forever' firm; the payments duopoly and UNH gone; an airline bought | Pruning, not churning — trimming stakes that had run up or whose thesis had aged is discipline, not hyperactivity; the core (Apple, AmEx, Coke, BofA) is untouched |
| Abel is unproven as an allocator — a superb operator (he built BHE), but never the public capital-allocator; the equity book is now effectively his call | 'He has launched' — Buffett on Abel's first elephant (Taylor Morrison): 'Greg did that faster than I could have, smoother than I could have, and I never talked to the CEO' |
| $397B cash unspent — the pile keeps growing; the tiny $235M buyback says even management doesn't find its own stock cheap | Discipline, not paralysis — refusing to overpay in a dear market IS the Buffett method; the cash is dry powder, and buybacks resumed after a 21-month pause |
| The double act is gone — no Munger to say 'no,' no Buffett to write the letter; the culture's two pillars both departed within ~2 years | The culture was built to outlast them — decentralised, trust-based, no committees; Abel's first letter reaffirmed 'disciplined investing into perpetuity' |
Our read: the changes look more like disciplined pruning than a break with the gospel — and, tellingly, several of Abel's sells mirror our own board's cautions (we flagged Visa and Mastercard as flawless businesses at full prices, and UnitedHealth as a fat pitch that had already been thrown). Selling a wonderful business that has become fully priced is not a betrayal of Buffett; it is Buffett, at his most disciplined. The Alphabet and Delta moves are the genuine departures, and they deserve a few years of scrutiny before judgment. But the honest answer to "is it still Buffett's Berkshire?" is: the machine is, the culture appears to be, and the jury on the allocator will be out for at least three years. You are buying a superb fortress whose single most important input — capital-allocation judgment — is mid-handover from the greatest practitioner in history to a capable, disciplined, but unproven successor. That is not a reason to avoid it. It is the reason it is not a table-pounding buy at 1.48× book.
It competes with the index, not with a company
| Arena | Who | Where Berkshire stands |
|---|---|---|
| As an investment vehicle | The S&P 500 index | 60-yr: 19.7% vs 10.5% — but the last decade merely matched it |
| For whole companies | Private equity (KKR, Apollo…) | Berkshire's permanent capital + no-flip promise wins the sellers PE can't |
| Insurance / float | Progressive · Chubb · reinsurers | Unmatched scale + a balance sheet that lets it write the biggest cat risks |
| Railroads | Union Pacific | Half of a Western-US duopoly — a wonderful, if capital-hungry, business |
| For the marginal dollar | Its own $397B in T-bills | The real rival is inertia — cash earning ~4% while it waits |
Berkshire's true competitor is not another company — it is the index fund, and this is the uncomfortable truth an honest analysis must state. For 1965–2015, Berkshire beat the S&P 500 in essentially every rolling period, a record without parallel. But over the most recent decade (2016–2025), Berkshire returned ~14.3% a year against the S&P's ~14.8% — the index, modestly, ahead. Buffett himself has said for years that Berkshire's size means shareholders should expect only "slightly better than the average American company," and the numbers now agree. The bull case is no longer "it will crush the index"; it is subtler and, for many investors, better: Berkshire offers index-like returns with far less risk — no debt, $397B of crash-insurance, a diversified base of real cash-generating businesses, and a call option on Abel deploying that mountain of cash brilliantly in the next dislocation. You are not buying outperformance; you are buying safety with upside optionality. Against private equity, meanwhile, Berkshire's advantages are structural and permanent: it never has to sell, never charges a fee, and is the only buyer a family founder trusts not to gut their company. In its two real businesses — insurance and owning-companies-forever — it has no equal.
The greatest record ever recorded — handing over the pen
This is the dial that matters most for Berkshire, and the one in the most flux. For sixty years this section would have needed a single word: Buffett. Now it needs a cast.
Read the capital allocation honestly and it cuts both ways. On one hand, this is the most disciplined steward in corporate history and its heir is behaving exactly as taught — refusing to overpay (for acquisitions, for its own stock, for anything), hoarding dry powder for the next dislocation, buying one sensible elephant (Taylor Morrison) when the price was right. That discipline is a feature, not a bug; most CEOs would have blown the $397B on empire-building years ago. On the other hand, the pile keeps growing because the world offers nothing big and cheap enough to absorb it, and cash earning 4% is a quiet tax on your return — the very "cash drag" that has helped the index catch Berkshire this past decade. And the tiny buyback is management telling you, in the plainest language they have, that they don't find their own shares cheap here. I score management an 8: the greatest legacy in the business, genuine discipline from the successor, marked down for the unproven allocator and the growing drag of un-deployed cash — a mark that should rise if Abel deploys well, and fall if the pile keeps swelling.
Modest reported returns, an unbreakable balance sheet
| Metric | Value | Read |
|---|---|---|
| 60-year record (1965–2025) | 19.7% vs S&P 10.5%/yr | ▲ the greatest long-run record ever recorded |
| Recent decade (2016–2025) | ~14.3% vs S&P ~14.8%/yr | ◆ honestly — the index modestly AHEAD |
| Cash & Treasury bills (Q1 2026) | ~$397B | ▲ ~38% of the company — fortress + drag |
| Insurance float | ~$176B | ▲ interest-free (often paid) investable capital |
| Book value (equity) | ~$727B | ▲ grows ~10%+/yr — the truest scorecard |
| Return on equity (TTM) | ~10.3% | ◆ modest & lumpy — mark-to-market distorts it |
| Total debt / equity | ~0.20 | ▲ almost no leverage — by design |
| Operating earnings (FY2025) | ~$44.5B | ▲ +~18% in Q1 2026 — the real engine hums |
| Dividend | $0 (paid once, 1967) | ◆ all reinvested — a compounding, not income, story |
Two numbers tell the whole story, and they point in opposite directions. The first is the 60-year record — 19.7% a year, a 6-million-percent cumulative return — the single greatest feat of compounding in the history of capitalism, and the reason Berkshire commands its premium and its reverence. The second is the recent decade — ~14.3% versus the index's ~14.8% — which says, without flinching, that the age of crushing the market is over, a victim of size. Reported return-on-equity of ~10% looks pedestrian and is genuinely distorted (accounting rules now force Berkshire to run unrealised stock gains and losses through net income, so the bottom line lurches by billions on nothing), which is why Buffett always told shareholders to watch book value and operating earnings, not net income — and both of those grow steadily and unglamorously at ~10%+. The one number with no asterisk is the balance sheet: $397 billion of cash, ~$176 billion of float, almost no debt. Whatever else 2026 holds, this is the single most unbreakable large enterprise on the planet — a company that cannot be forced to do anything by anyone, ever.
Buffett's own two-column method · fair, not cheap
| Yardstick | Today | Context | Read |
|---|---|---|---|
| Price / book value | ~1.48x | post-2000 range ~1.2–1.6× · buyback zone historically <1.35× | upper half of its own range |
| P/E — trailing | ~14.9x | flattered by mark-to-market gains | optically cheap, but distorted |
| Price / operating earnings | ~24x | on ~$44.5B operating earnings | the truer multiple — full, not cheap |
| Graham Number | ~$505 | ≈ today's price | priced right at 'defensive fair value' |
| Management's buyback signal | $235M (token) | vs tens of billions when they think it's cheap | insiders say: not a bargain |
Berkshire is genuinely hard to value, so Buffett handed us the method himself: the two-column approach from his 2010–11 letters. Column one is the investments per share — the cash, bonds and stock portfolio, roughly $385 a share, assets you can mark to market today. Column two is the operating businesses — BNSF, BHE, insurance underwriting, the manufacturing empire — earning ~$44.5B pre-tax, which you capitalise at a fair multiple and add on top. Do the arithmetic and you land close to where the stock trades: Berkshire at ~1.48× book is fair value — neither the bargain it was at 1.2× in past panics, nor egregiously dear. Three independent signals agree it is fully priced: the price-to-book sits in the upper half of its 25-year range; the Graham Number (~$505) sits right on top of the price; and — the loudest signal of all — management bought back only $235 million of stock, a rounding error against a $1.08 trillion company, which is Abel telling you in the clearest language available that he does not consider his own shares cheap here. (When Berkshire does think it's cheap, as in 2020–21, it repurchases tens of billions.) Even the thin analyst coverage sees ~7% downside to $465. None of this makes Berkshire a sell — a fortress compounding book value at ~10% with a call option on the next crisis is a fine thing to own at fair value. It simply means there is no margin of safety at $502, and the disciplined move is the one Buffett himself models: own it, and let the price come to you. The buy zone opens toward ~1.35× book (~$455) — near the 52-week low, and the level where Berkshire's own buyback appetite has historically switched on.
The succession, the cash, the wildfire · verified July 2026
Verified the week of publication. The ruby risk is not financial — it is the succession itself: a 95-year-old chairman, an unproven (if disciplined) allocator, and a $1.08 trillion machine whose entire historical edge was one man's judgment. Everything else is amber. On litigation, Berkshire's biggest tail is the PacifiCorp wildfire liability (its BHE utility): a June 2023 Oregon jury found the utility grossly negligent for the 2020 fires; Berkshire has accrued $2.853B cumulative and paid $2.276B, with ~$577M remaining. Crucially, on 8 April 2026 the Oregon Court of Appeals reversed and remanded the class-action verdict (CNBC: "could save it billions") — materially de-risking the old "tens of billions" fear, though plaintiffs have petitioned the state Supreme Court, so it is reduced, not resolved. The HomeServices real-estate antitrust saga settled for $250M (≈$100M still outstanding on appeal, plus two Texas cases). Note what is absent: no securities fraud, no accounting scandal, no regulatory assault on the parent — for a firm this size and age, an astonishingly clean record, which is itself a statement about the culture. The amber financial risks are structural and permanent: the cash drag, the size curse that has let the index catch up, the conglomerate discount, and the succession-behind-the-succession (Jain is 74). None threatens solvency — this is the most unbreakable balance sheet in the world. The only real question is return, and return now rests on Abel.
I have put off writing this letter for as long as I decently could, and I should tell you why. Every judgment we render on this website — every time we ask whether a business is wonderful or merely fair, whether a moat will hold, whether a price offers a margin of safety, whether a management deserves the capital it commands — we render in a borrowed voice, and the voice is Warren Buffett's. To turn that lens on Buffett's own company felt like a student grading his teacher's exam. But the reason I can no longer avoid it is the same reason it has become the most important thing on our board: for the first time in sixty years, the exam is being written by someone else. Warren Buffett is ninety-five and remains Chairman, but on the first day of this year Greg Abel became the man who decides where Berkshire's next dollar goes. The teacher has handed over the pen.
Let me first say plainly what Berkshire is, because familiarity has dulled how astonishing it remains. It is the single greatest act of compounding in the recorded history of capitalism. One dollar entrusted to Buffett in 1965 became roughly sixty-one thousand dollars by the end of last year — a return of six million percent, against thirty-nine thousand percent for the market, nearly double the index every year for six decades. No one else has ever done anything remotely like it, and no one else ever will, because no one else will get sixty years and a machine this well-built to do it with. Underneath that record sits a fortress: three hundred and ninety-seven billion dollars in cash, a hundred and seventy-six billion of insurance float that costs less than nothing to hold, almost no debt, and a diversified base of real businesses — a railroad, a utility empire, insurers, See's Candies, a phone-maker's worth of Apple stock — that between them own a slice of the entire American economy. It is the most unbreakable large enterprise on the face of the Earth. Nothing and no one can force Berkshire to do anything, ever. In a world of leverage and fragility, that is worth more than any dividend.
And I have to tell you, with the same honesty Buffett always used on himself, two things that temper the awe. The first is that the age of beating the market is, by the arithmetic of size, over. Over the last ten years Berkshire compounded at about fourteen percent a year and the index compounded at about fifteen — the pupil has, for a decade, been quietly overtaken by the average of American business, exactly as Buffett warned would happen once the snowball grew too large for any hill. You are no longer buying outperformance. You are buying something subtler and, for many, better: roughly index-like returns with a fraction of the risk, plus a three-hundred-and-ninety-seven-billion-dollar call option on the next time the world panics and Berkshire, alone and unafraid, gets to be the buyer. The second is that the one input that made all the magic — a single man's judgment about where to move each dollar — is precisely the input now changing hands. Greg Abel is a superb operator and, on the early evidence, a disciplined allocator: his first great deal earned Buffett's own benediction, "he has launched"; his portfolio pruning, selling the fully-priced Visas and Mastercards and UnitedHealths, looks far more like Buffett-at-his-most-disciplined than like a break with the faith. But he is unproven at the one thing that matters most, and he will be for at least another three years. That is not a reason to stay away. It is the reason this is not a stock to pound the table on.
Which brings me to the price, and here Berkshire is refreshingly free of the agony that Costco or Visa put us through. It is not expensive and it is not cheap. At about one-and-a-half times book value it sits in the upper half of its own twenty-five-year range; Benjamin Graham's old defensive-value formula lands almost exactly on today's quote; the thin Wall Street coverage sees a few percent of downside. And the loudest voice of all is management's own: in the first quarter of this year Berkshire bought back a mere two hundred and thirty-five million dollars of its own stock — a rounding error against a trillion-dollar company — which is Greg Abel telling you, in the only language a disciplined buyer of businesses ever fully trusts, that he does not think his shares are a bargain here. When Berkshire truly believes it is cheap, as in 2020 and 2021, it repurchases tens of billions. It is not doing that now. Believe him.
So here is my verdict, and it is the one I would give a member of my own family who asked what to do with a piece of the greatest business ever built. Own the fortress — and add on weakness. If you own Berkshire, this is not a moment to sell a compounding, fortress-balance-sheet machine that grows its intrinsic value at ten percent a year and holds the largest war chest on the planet for the next dislocation; you sell that for taxes and regret. But if you are buying, understand that at five hundred and two dollars there is no margin of safety — you are paying fair value for fair value, in the middle of the most consequential management transition in corporate history. The disciplined thing, the Berkshire thing, is to do what Buffett taught: refuse to overpay, keep your powder dry, set your price, and let the market bring the stock to you. Berkshire itself will show you where — its own buyback appetite switches on toward one-and-a-third times book, around four hundred and fifty-five dollars, near this year's low. That is the level at which you are paid to take on the one uncertainty that matters, which is Greg Abel. Buy it there, hold it forever, and let the fortress do what it has always done. The pen has changed hands. The machine has not. And for the first time, we get to end a letter not by quoting the man, but by watching whether the house he built can write in his hand without him.