Berkshire's First Buyback in Three Years Was Not Buffett's Call

Berkshire Hathaway ended March 31 with roughly $400 billion in the till and June 30 with $365.5 billion.

The $34.5 billion difference is the first meaningful drawdown in the company's sixty-year history that Warren Buffett did not authorize.

The man who did is Greg Abel, chief executive since January 2026; Buffett stayed on as chairman.

When Berkshire published second-quarter results on August 8, the question the market had been asking for months got numbers: what will the new management do with the money?

How the Pile Got This Big

Cash accumulation was never a strategy but the residue of an absence. Buffett sold stock — most of Apple included — replaced it with nothing and stopped buying back Berkshire's shares.

The repurchase rule is a measuring stick, not a commitment: buy below intrinsic value, not above.

For three years cash grew in the shadow of that rule: today's $365.5 billion is roughly a third of the company's $1.1 trillion market capitalization.

Abel restarted it in March, saying the stock traded below intrinsic value. First-quarter purchases came to $234 million, the second quarter's to $4.5 billion; the six-month total is $4.8 billion.

The Trouble With "Profit Doubled"

Most of Saturday's headlines led with net earnings doubling, $12.4 billion to $25.7 billion. That is true, and it tells you almost nothing about the business.

An accounting rule in force since 2018 pushes daily portfolio price swings into the income statement. A few percent there can outweigh what the railroad and the utilities earn in a quarter.

Operating earnings carry an asterisk too: about $1.2 billion of the increase came from currency moves on yen- and euro-denominated debt. Strip it out and the underlying businesses grew around 5%, not 16%.

The Mechanism: When a Buyback Creates Value

A repurchase is a company buying and cancelling its own shares. Whether holders gain or lose depends entirely on the price paid.

At Berkshire's scale, $4.5 billion is four-tenths of one percent of a $1.1 trillion market value; even 10% below intrinsic value, the gain to holders is hundredths of a percent.

Management buys only when it thinks the stock is cheap, so the $4.5 billion says what Abel thinks the shares are worth. So, for three years, did the silence.

In a valuation argument, management's own money outweighs its presentation.

The pile sits largely in short-dated Treasury bills, where every one-percentage-point move in short-term rates is worth roughly $3.7 billion a year on $365.5 billion — with no decision taken.

The effect showed up this quarter: insurance investment income fell 9.1% to $3.1 billion on lower short-term rates.

The macro backdrop is softening too: July nonfarm payrolls fell by 23,000 against expectations of an 86,000 gain, with May and June revised down by a combined 103,000. The 10-year Treasury yield slipped to 4.60%. More: Interest Rates and Bonds and Employment Data

Where the Money Went

The quarter's biggest move landed on June 1. Alphabet announced an $80 billion equity raise for AI infrastructure. Of that, $30 billion came in underwritten offerings, $40 billion at-the-market from the third quarter, and $10 billion was placed directly with Berkshire.

Berkshire took $5 billion of Class A at $351.81 and $5 billion of Class C at $348.20; per Forbes, Buffett negotiated it himself.

Alphabet guides to $180–190 billion of capex for 2026 and says 2027 will rise significantly. Pichai: "The risk of under-investing is dramatically greater than the risk of over-investing."

The Alphabet position appeared in the third quarter of 2025, tripled in Abel's first quarter and grew again in June. At roughly $41 billion it is the fourth-largest holding, behind only Apple and American Express.

In that same first quarter the Visa, Mastercard, Domino's Pizza and Amazon positions were closed and about $2.6 billion went into Delta.

What the Businesses Did

Insurance is the weak link: GEICO's loss ratio deteriorated roughly five points to 76.6% on higher claim frequency, 10–12% higher bodily-injury severity and heavier commission and advertising costs. Pretax underwriting earnings fell from $1.821 billion to $994 million.

BNSF moved 6.5% more carloads and grew earnings 6.3% despite a 68% jump in fuel expense; energy earnings rose 26.9% on a 3.1% increase in retail electricity volumes.

Berkshire's own shares lagged: through August 7 the stock was up 3.5% for the year against 14.1% for the S&P 500.

The Other Side

What Remains

The real question is still open: will $365 billion turn into an acquisition large enough to matter? The $34.5 billion spent in a quarter shrank the pile by only 8.6%.

A second clock runs on ownership. In July Buffett moved roughly $6 billion of Class B stock to four family foundations. He said he aims to distribute all his Berkshire shares within about eight years.

Buffett's shadow balances Abel's decisions today; that shadow shrinks with the calendar.

This piece draws on Berkshire Hathaway's second-quarter results of August 8, 2026, Alphabet's June 1, 2026 equity-raise announcement, and reporting on the quarter from Forbes, Fortune and the Associated Press. Segment comparisons and the separation of currency-driven gains come from independent analyst notes and may not map exactly onto the company's official segment definitions.

Correction (September 13, 2026): prose tightened for rhythm; figures and sources unchanged.