The same idea in one frame: long stretches of accumulation, short windows of deployment, and a present position that has not yet met its window.
Berkshire Hathaway is holding a record $397 billion in cash and Treasury bills. That is usually reported as a warning. It is better understood as a purchase order waiting for a price.
Defense and opportunity are not opposites. In every serious market reset of the last ninety years, the investors who bought the wreckage were the same people who had spent the preceding years looking too cautious, too early, and too slow.
When a large investor holds a large cash position, the financial press says he is sitting on the sidelines. The phrase is borrowed from sport, where the sideline is where you stand when you are not allowed to play. It is the wrong metaphor, and the wrong metaphor has cost ordinary investors a great deal of money.
Cash is the only asset that converts into any other asset at a moment of the holder's choosing. Its worth is not the yield printed on the Treasury bill. Its worth is the price of everything else on the day you decide to spend it. Held for its own sake, cash is a slowly depreciating asset. Held against a specific future opportunity, it is a call option on someone else's forced selling — and unlike a listed option, nobody charges you a premium for it. You pay in a subtler currency: the return you did not earn while you waited.
That is the whole trade. Defense and opportunity are not opposites. They are the same position observed at two different moments in a cycle. The difficulty is that the costs arrive first, quarterly, visibly, in the form of performance you can watch your neighbor collecting. The payoff arrives once, compressed into weeks, and only if you are willing to act at the exact moment acting feels insane.
What follows is the record: who has run this trade, what it cost them while they waited, what it returned when the window opened — and the honest, serious case against trying to run it yourself.
Between 2022 and 2024, Berkshire sold a net $172.9 billion of equities, with $134.1 billion of that in 2024 alone. Apple was cut from nearly half the equity portfolio to roughly a fifth. Bank of America was reduced by more than half. Buybacks stopped for twenty-one consecutive months because Buffett judged his own stock too expensive to repurchase.
Greg Abel took over as chief executive at the start of 2026, and the first question every shareholder asked was whether the new man would spend it. In his first quarter, Berkshire sold another $24.1 billion of stock against $16 billion of purchases and the cash pile rose to its record. Whatever else that is, it is not drift. Two different people, running the same balance sheet, reached the same conclusion about price.
First, Berkshire is not literally on the sidelines. Abel's first major transaction was the $6.8 billion acquisition of homebuilder Taylor Morrison at $72.50 a share, a 24% premium, and buybacks resumed in March 2026. The company is not refusing to invest. It is refusing to overpay, which is a different behavior with a different meaning.
Second, the cash is partly a function of size. A company with $397 billion to place cannot buy a good small business and move the needle; it needs elephants, and elephants are scarce at these valuations. Your household does not have that constraint. What is prudence for Berkshire may simply be paralysis for you.
Third, and most important: this position has cost real money. Over the twelve months to the Q1 2026 report, Berkshire's stock fell about 11% while the S&P 500 gained roughly 29%. That is a forty-point gap, paid in full, in public, by the most respected capital allocators alive. Anyone who tells you defense is free is selling something.
In each case below, the same structure repeats. A long, unglamorous period of accumulation while the crowd compounds faster. A short, violent window when assets change hands at prices that make no sense to anyone who has to sell. And a payoff that arrives years later, in the form of positions bought at levels that were never available again.
Templeton arrived on Wall Street in 1938, into the wreckage of the Depression, at a moment when American equities had spent a decade teaching investors that stocks destroy families. Rather than deploy into the recovery narrative, he kept borrowing capacity in reserve and waited for a price he considered absurd.
September 1939. As war broke out in Europe, he borrowed $10,000 and bought 100 shares in each of 104 companies trading under a dollar on the New York Stock Exchange — including 34 already in bankruptcy. Four went to zero. He made money on the other hundred and sold roughly four years later for about $40,000.
At the top of the late-1960s speculative market, Buffett did something no fund manager does: he wound up the Buffett Partnership and handed the money back, telling his partners he could no longer find ideas in a market he understood. He then sat, largely in cash and bonds, through the Nifty Fifty mania while others compounded around him.
The 1973–74 bear market cut the S&P roughly in half. In 1973 Berkshire bought The Washington Post — a business Buffett judged worth several times the price the market was assigning it. Berkshire held the position for four decades; by the time it was exchanged, the stake was worth more than a billion dollars.
On December 27, 1999, Barron's ran a cover story asking what was wrong with Warren Buffett and suggesting he might be losing his touch. Berkshire's Class A shares were down sharply on the year while the S&P rose about 20%, and Berkshire's entire market value was smaller than Yahoo's. He was described, in print, as too conservative and possibly finished.
The S&P 500 then fell in 2000, 2001 and 2002. Berkshire rose more than 26% in 2000 while the index lost about 9%, gained 6.5% in 2001 against an index down roughly 12%, and slipped less than 4% in 2002 while the index fell 22%. Yahoo was eventually sold for a fraction of its 1999 value.
Between January 2007 and May 2008, in anticipation of a downturn that had not yet arrived, Howard Marks and Bruce Karsh raised $14.5 billion across two distressed debt funds — including $10.9 billion for a single vehicle, at that point the largest distressed fund ever raised. For more than a year it mostly sat there while credit markets kept rallying.
Lehman Brothers failed on September 15, 2008. In the fifteen weeks that followed, Oaktree invested more than $5.3 billion — over half the fund's eventual drawn capital — into bank debt and securities at distressed prices. The fund reported a net internal rate of return above 31% from inception through the end of 2009.
Through the credit boom of 2003–2007, Berkshire's cash accumulated while private equity and structured credit competed away every attractive return. Buffett wrote repeatedly that he could not find prices he liked. He was accused, again, of missing the cycle.
In roughly ten days of September and October 2008: $5 billion into Goldman Sachs preferred stock at a 10% dividend with warrants struck at $115, and $3 billion into General Electric on similar terms with warrants at $22.25. The Goldman position ultimately produced about $3.7 billion. In August 2011 he did it again — $5 billion into Bank of America at 6%, with warrants on 700 million shares at $7.14, later exercised to make Berkshire the bank's largest shareholder.
The reason these episodes compress into weeks is mechanical, not psychological. Prices collapse furthest when holders are selling for reasons that have nothing to do with value: margin calls, fund redemptions, covenant breaches, insurance reserve requirements, a fiscal-year-end that will not move. Forced sellers are price-insensitive by definition. Once their positions have cleared, the marginal seller becomes someone who is merely pessimistic — and pessimists hold out for better prices. The bargain disappears well before the news improves.
Ray Dalio has framed the current version of this risk precisely: bubbles tend to burst not when the technology disappoints but when holders of paper wealth are forced to convert it into spendable money — for debt service, tax bills, or redemptions — and discover that the wealth was never money. Bridgewater estimates that four large technology companies alone may spend around $650 billion on AI infrastructure in 2026, against roughly $410 billion in 2025. Spending of that magnitude has to be financed, and financing is where a valuation story meets a cash flow.
Berkshire is the largest example, not the only one. The people below manage money in very different ways and agree on very little, which is what makes the overlap worth noting.
A record $397.4 billion in cash and Treasury bills at the end of Q1 2026, roughly 59% of the investable portfolio and more than a third of the company's market value. Net seller of equities again in the quarter under a new chief executive.
Has spent 2025 and 2026 arguing that U.S. equity valuations are elevated and that investors should avoid the sectors the market loves most, where high prices and heavy leverage tend to arrive together. His firm's entire business model is holding capacity for dislocations that have not happened yet.
Told the 2026 Global Alts conference that he cannot build conviction about businesses priced at forty times earnings — or at no meaningful multiple at all — because doing so requires certainty about a very distant future. Baupost's disclosed equity book is small, concentrated in durable cash generators, and paired with a substantial cash allocation and an interest in distressed credit.
Describes the AI complex as the largest bubble in American history and has said the most expensive names could fall as much as 70%. He recommends shifting toward international equities, bonds and precious metals. He is speaking personally rather than for GMO, and he has been early before — by years.
Says his bubble indicators now sit near levels last seen in 1929 and 2000, and points to a vulnerable window in the political calendar. Notably, he pairs the warning with advice against panic selling — his counsel is to expect lower forward returns, not to evacuate.
American money market funds held $7.86 trillion in the week ended July 22, 2026, off a record $7.95 trillion set weeks earlier. Retail money accounts for $3.08 trillion of it. Caution is not a contrarian position at the moment. It is a very crowded one.
A page that only made the bullish case for caution would be advertising, not analysis. These objections are strong. Two of them I think are decisive against the naive version of this strategy.
The evidence that investors who move to cash and back underperform buy-and-hold is overwhelming and not seriously disputed. Missing a handful of the best days — which cluster inside the worst months — does most of the damage. Nothing on this page argues for exiting the market on a valuation reading. The distinction that matters is between a reserve sized to your obligations and a bet sized to your opinion.
Buffett waited four years between closing the partnership and the 1973–74 bargains. Grantham was early on the last two bubbles by margins that cost GMO a large share of its client base. Someone who moved to cash on Marks's valuation warnings in August 2025 watched the S&P index fund climb roughly 19% over the following nine months. Being right eventually and being solvent, employed and psychologically intact throughout are different achievements.
In February and March of 2020, the fastest crash in modern history arrived, Berkshire had tens of billions available — and Buffett largely did not deploy it. He sold the airlines instead and was criticized, fairly, for it. The window opened and the most disciplined buyer alive did not step through. Liquidity is a necessary condition for opportunism. It is nowhere near a sufficient one.
Treasury bills currently pay something. After tax at ordinary income rates and after inflation, the real return on a large cash position is thin and can be negative. Hold it for a decade waiting for a reset that arrives in year eleven and the arithmetic can eat much of the advantage you were waiting to capture.
A CAPE reading above 40 tells you a great deal about likely returns over the next decade and almost nothing about the next twelve months. The ratio crossed 40 in 1998 and the market rose substantially for two more years. Every valuation metric on this page should be read as a statement about expected return, not about timing.
In thirty years of watching households do this, the pattern is consistent. People de-risk out of fear, then wait for confirmation that the danger has passed — and confirmation arrives only after prices have recovered. They sell low, buy back higher, and pay for the same decision twice. If you cannot state in advance, in writing, what would cause you to redeploy, you do not have a defensive strategy. You have an exit.
Berkshire's cash and your cash are not the same instrument, and confusing them is how thoughtful people end up making expensive mistakes. Berkshire's reserve exists to buy assets cheaply. Yours exists so that you are never the person selling assets cheaply. Those are related, but they are not identical, and only one of them requires you to be right about the market.
This distinction matters most in households like the ones I work with: a career at one utility, a pension election coming up, health coverage tied to the same employer, and a retirement account with an uncomfortable share of it sitting in that employer's stock. In that structure, a market reset is not an abstraction about valuations. It is a question of whether you are forced to sell into it.
That last line is the point of the page. For an institution, dry powder is about buying the bottom. For a household within a decade of retirement, it is about not being part of what makes the bottom. Those require completely different amounts of foresight — and only the second one is achievable on purpose.
I have written for years about the Federal Reserve, about the debt, and about a political class in both parties that has treated compounding obligations as somebody else's problem. I have not changed that view, and the numbers since have not argued me out of it. When the national debt was around eighteen trillion dollars, serious people were already calling it unsustainable. It is now approaching thirty-nine trillion. Nothing about the intervening decade suggests the problem was solved. It was postponed, and postponement has a price that gets paid in one lump.
So yes — I think the largest financial reset of my lifetime is ahead of us rather than behind us. What I will not do is pretend that conviction gives me a date. It does not. Being directionally right about a decade and specifically wrong about a quarter is a good way to destroy a retirement, and I have watched it happen to careful people.
What that view actually changes in practice is narrow and unglamorous. It makes me less willing to let a single employer's stock stay at forty percent of a portfolio because the cost basis is low and the dividend feels safe. It makes me insist that the first several years of retirement spending sit somewhere that a bear market cannot reach. It makes me want redeployment rules written down in advance, while the client is calm, because nobody writes good rules in March of a crash. And it makes me deeply skeptical of any plan whose success requires the last fifteen years to keep repeating.
Not to be answered in your head, where the answers move around. In writing, dated, where you can be held to them by your future self in a much less comfortable market.
If the market fell 40% starting next month and stayed down for three years, what would I be forced to sell, and in what month would I have to sell it?
How many years of my actual spending could I fund without touching a single share of stock? Not a percentage — a number of years.
What share of my household's total position — salary, pension, health coverage, and portfolio — depends on the fortunes of one company?
If I hold cash and prices do fall, what specifically would trigger me to put it back to work? A price level, a valuation, a calendar date, a schedule? If I cannot name it now, I will not find it then.
If I am wrong and the next ten years look like the last ten, does my plan still get me where I need to go — or does it only work if I get the reset I expect?
Every episode on this page has the same shape: the position that mattered was taken while nothing was happening. Oaktree could not have raised that fund in October 2008. Buffett could not have found $8 billion in the week Lehman failed if he had not spent five years declining to spend it. The decision that looks brilliant in the crisis was made in the boredom that preceded it.
If you are inside ten years of retiring from a utility career, the useful version of that work is specific and finite: know what you would be forced to sell, know how concentrated you actually are, and decide your rules while the market is calm enough to let you think.
Berkshire Hathaway Q1 2026 earnings release and 10-Q; Berkshire 10-K filings via SEC EDGAR. Goldman Sachs 8-K of September 23, 2008 and FY2008 10-K for the terms of the preferred issuance and warrants. Oaktree Capital Group Form S-1 for the 2007–2008 fundraising and post-Lehman deployment figures; Institutional Investor for the reported fund IRR. Barron's, December 27, 1999. Investment Company Institute weekly money market fund assets, release of July 23, 2026. Shiller CAPE data via Robert Shiller's series. Contemporaneous reporting from Bloomberg, Reuters and CNBC on the 2008–2011 Berkshire transactions, the 2026 Global Alts conference remarks, and public comments by Ray Dalio and Jeremy Grantham. Historical Templeton figures from published accounts of his 1939 purchases.