Warren Buffett says Berkshire Hathaway's Alphabet bet was his call, and the timing says something useful about the AI race: he likes Google more as it starts to look less like a software company and more like an infrastructure owner.
Warren Buffett has finally answered the Alphabet question. In a July 15 interview with CNBC's Becky Quick on Squawk Box, he said Berkshire Hathaway's move into Google's parent company was not simply Greg Abel's first big mark on the portfolio. Buffett made the call. He also made clear that Abel is now the one who has to live with those calls, as Berkshire's chief executive.
That distinction matters. It tells you Berkshire's Alphabet stake is not a stray tech trade dressed up as succession planning. It's Buffett admitting, late but plainly, that Google has become too profitable and too central to ignore. According to a Reuters summary of the CNBC interview, Buffett said he made a mistake by not investing in Google earlier and now sees Alphabet as more likely to be a winner based on its record.
The numbers are large enough to deserve less romance. Reuters reported in November 2025 that Berkshire first disclosed 17.85 million Alphabet shares as of September 30, a stake worth about $4.3 billion at the time. Alphabet's own SEC filing then showed that on June 1, 2026, it agreed to sell Berkshire 14,212,035 Class A shares at about $351.81 each and 14,359,656 Class C shares at about $348.20 each. That raised $10 billion from Berkshire through a private placement. CNBC, through Reuters, put Berkshire's Alphabet holding at roughly $31 billion this week.
That's real money. It is also still not Buffett surrendering to every AI story on Wall Street.
Google now spends like a railroad
Buffett's old objection to technology was never really about the money - it was about the moat, and whether he could judge how long it would hold. Google Search, YouTube, Android and Google Cloud settled that question years ago. Then came AI. It upended Alphabet's whole spending profile. The company is no longer only an advertising machine with code on top. It is buying chips, land, power and data center capacity at a pace that looks closer to a utility buildout than a web business.
Alphabet said in a June investor presentation that it expects 2026 capital expenditures of $180 billion to $190 billion - about six times its 2022 level. Most of that is going into technical infrastructure. That's a staggering jump. The company initially announced an $80 billion equity raise to fund the AI buildout, then said in a June 2 SEC filing that the raise had been upsized to $84.75 billion. That included common stock, mandatory convertible preferred shares, an at-the-market program and Berkshire's $10 billion private placement.
You don't need to love every AI projection to understand why Buffett noticed. Berkshire owns BNSF Railway and Berkshire Hathaway Energy, businesses where capital goes into the ground for years before the payoff becomes visible. Track, substations, transmission lines, data centers, chips. The assets are different, but the discipline is familiar.
Here is the thing: Alphabet is spending like that because it has to, not because the economics are already settled.
The bet is on endurance
Buffett's comments landed inside a bigger argument: can the largest tech companies keep pouring money into AI infrastructure without wrecking their margins? Bloomberg reported in April that the biggest US tech firms planned as much as $725 billion of capital spending this year, mostly for AI data center equipment. That's not a typo. Morgan Stanley figures cited by Axios put hyperscaler spending above $750 billion this year, and potentially far higher by 2030. That's the scale of the bet.
That is the uncomfortable part of the story. Alphabet, Microsoft, Amazon and Meta can't easily step back if customers keep demanding compute and rivals keep building capacity. Pull back too soon and you risk handing the next platform shift to someone else. Keep spending and you risk finding out that the revenue curve is slower than the construction bill.
Buffett is not endorsing the whole race. He is choosing one runner. Alphabet still has the search and advertising engine, YouTube, Cloud momentum and a balance sheet strong enough to fund a brutal investment cycle without looking desperate. If you are going to own one company forced into an AI spending war, you want the one that can absorb a bad year and still come back with cash.
There is a useful restraint in Buffett's answer too. Reuters said he noted Abel is the decider and that neither man does anything the other disapproves of. That is Berkshire's transition in one sentence: Buffett's fingerprints are still on the portfolio, but Abel owns the next chapter.
Alphabet is not suddenly a classic Buffett stock because it owns data centers. It is a classic Buffett problem because price and cash flow matter now, and so do capital discipline and competitive endurance - more than the slogan attached to AI. Frankly, that is a better test than asking who has the flashiest model this quarter.
Also read: TSMC Pledges Another $100 Billion for US Chip Plants After Record Quarter, Kioxia's Market Value Has Nearly Halved Since Its AI-Fueled Peak in June and A Regulatory Filing Reveals DeepSeek Is Now Worth About $52 Billion
