Berkshire's $38 Billion Google Bet Is Real. Here's the Part You Can't Copy

The headline is easy to repeat: Berkshire Hathaway just packed roughly $38 billion into Google's parent, making Alphabet its third-largest stock behind only Apple and American Express. On its face it reads like the world's most famous value investor handing the AI era its biggest endorsement yet. Headlines skip the part that tells you the most, though, which is how the money moved. Roughly half of that new stake did not come from Berkshire buying shares in the open market. It came from Google selling Berkshire $10 billion of brand-new shares at a 6.5% discount to the market price.

That single check was the anchor in an even bigger event: Alphabet's roughly $85 billion equity raise, the largest equity offering in history, which it used to fund the AI buildout. Pause on what that means. Google is one of the great cash machines in the history of business — the profitable engine of search and ads that has funded moonshots for two decades. And it still needed to go to shareholders and sell $85 billion of stock to keep its data centers multiplying. This is the AI boom colliding with capital: a business that generates enormous cash, asking the market to hand it even more.

The person now closest to the money at Berkshire is CEO Greg Abel, and on Wednesday he gave the fullest public explanation of the trade yet. His argument is worth taking seriously because it comes from the demand side. Berkshire runs dozens of operating companies — insurance, energy, railroads, manufacturing, retail — and Abel says it has "a lot of visibility from within our companies as to how we're using AI, what type of benefits it's delivering". In other words, Berkshire is not guessing that AI pays. It is watching dozens of its own businesses deploy the technology and seeing results. On that evidence, he calls Alphabet a "significant player" in AI. That is the exact kind of rent evidence a disciplined buyer would want before writing a check this size: the user side confirming the product works.

It is worth being precise about who actually made the decision, because the story the outside world tells — "the new CEO is betting on Google" — is only half right. Warren Buffett, who handed the CEO role to Abel at the start of the year, says he initiated the position, and Abel calls him "the decider." And here the endorsement gets subtler: Buffett has said he ranks Alphabet below at least four or five of Berkshire's other holdings. So this is a real and large position — the third-largest stock in the portfolio — but it is not necessarily the highest-conviction one. Berkshire bought a lot of it, at a discount, in a company Buffett once admitted he made a mistake by not buying sooner.

The discount is the detail a retail investor should sit with, because it is the one part of the trade you cannot have. Berkshire got new Alphabet shares 6.5% cheaper than market on that $10 billion placement, plus roughly $7 billion more bought quietly on the open market during the quarter, bringing its stake up 83% to about 106 million shares. A buyer of that scale with inside knowledge of AI adoption, buying at a discount, is a materially better deal than buying Google shares at the market in your brokerage account. The validation is real information. The price is not something you inherit.

That distinction matters because of what Google is spending the money on. Alphabet guided 2026 capital spending as high as roughly $190 billion — more than double what it spent in 2025 — and it has raised that amount again since. In the second quarter, capital expenditures hit about $44.9 billion, roughly double the year-earlier period, and free cash flow turned negative by around $5.9 billion, the first negative reading since Google went public in 2004. Translation: the AI buildout is a bonfire of capital, and for the first time in its public life, Google is burning more cash than its operations throw off. It sold equity specifically to buy time and capacity before consumers — cash flow — caught up. The bet Berkshire is endorsing is not just "Google sells more ads." It is "Google can pour this much into compute and eventually earn it back, protecting a roughly 36% operating margin while Google Cloud grows near 50%."

Then there is price. Alphabet shares trade around $337, putting the company near $4.1 trillion in market value. By enterprise value to EBITDA, Google changes hands at roughly 23 times — richer than Microsoft (about 19), Meta (about 14), or Amazon (about 16). The market has not been punished for that optimism; the stock is up about 43% over the past year. In other words, the crowd already believes the AI story that Berkshire has now helped validate. The celebrated vote of confidence arrives with the votive already priced in.

So what actually changes for you? The Berkshire purchase is a genuine and useful data point: a demanding, cash-rich operator with visibility into real AI usage chose Google as its concentrated AI bet, and furnished Google the capital to stay in the race. That lifts the odds that Google's spend becomes earnings. But it is confirmation, not a shortcut. You cannot buy Berkshire's discount, Berkshire itself only ranks Alphabet below several other favorites, and the stock is priced for the buildout to keep working. The trade is not automatically yours to copy.

The number to keep watching is not Berkshire's next 13F. It is Google's free cash flow. If that turns positive again while Cloud keeps growing and the roughly 36% operating margin holds, the buildout is paying for itself and today's premium looks more defensible. The signal that the honeymoon is ending is the opposite one — capex keeps ratcheting up while cash flow stays negative, forcing more dilution to buy more capacity. Berkshire bought its shares at a discount and can afford to wait out that math at 6.5% off. Only one of those facts transfers to you.