Has Warren Buffet invested into Google for AI?

For decades, Warren Buffett kept Berkshire Hathaway away from big technology names. Then in 2025 the conglomerate disclosed a first-ever stake in Google's parent, Alphabet, and has kept adding to it. The third-quarter 2025 filing showed roughly 18 million shares worth about $4.3 billion.[1] By mid-2026 the position had grown many times over. It was valued at $36.6 billion, making it Berkshire's third-largest holding.[2] Buffett has said the idea was his. "I initiated it," he told CNBC in July 2026.[3] Four lines of reasoning explain why the bet makes sense for Berkshire.

1. A long-term holder who can look past the AI-bubble debate

Berkshire is not a trader of themes. Its stake grew through the AI spending debate, not after it. Berkshire bought another 36.4 million Class A shares in the first quarter of 2026 and 24.5 million more in the second.[4] Roughly $10 billion of that came directly from the company. About 60% of the new position came through a private placement tied to Alphabet's June capital raise.[5]

The worries are real. Alphabet expects capital expenditures of $195 billion to $205 billion this year, and investors fear the returns may take longer than expected or never arrive.[6] Buffett acknowledged the issue. He pointed to the enormous capital commitments needed to compete in AI as a key issue facing Alphabet and its rivals.[3] Berkshire still bought. For an investor with a multi-decade horizon, a stock that might swing sharply in the short term is a price paid for ownership, not a reason to stay out. So if any of the retail investors believes that Berkshire has bought Google for a year or two or three, it is not even remotely the case. They are likely to hold it for decades not years unless there is a change in Google’s business economics. 

2. A capital base so large that good options are scarce

Berkshire's problem is size. A great idea that moves the needle has to be very large. Berkshire held $365.5 billion in cash and equivalents, and its buybacks in the second quarter were made at around 1.4 times book value.[2] That is a price level Berkshire has used for buybacks before, and it suggests management does not see many cheaper alternatives. Berkshire also needs targets that can absorb billions without distorting the price. Alphabet is large enough that Berkshire's holding accounts for a little over 10% of its $358 billion stock portfolio.[4] Few companies can take that much capital. The long-running assumption that Berkshire is swimming in opportunities does not match the numbers.

3. Reasonable certainty on cash flow

Buffett's long-standing test is predictability. Alphabet's cash generation from its legacy advertising and cloud operations gives the company a reliable base to fund its AI buildout. Alphabet is also the parent of YouTube and a major player in cloud infrastructure.[7] Buffett knew the advertising engine firsthand. He has said he saw Google's ad business succeed through Geico, an early customer, but wasn't confident Google would emerge as the long-term winner.[3] He now treats that doubt as a mistake. He was candid that it is not his top pick: while a sizable holding, Alphabet wasn't among his very favorite businesses owned by Berkshire.[3] The stake is a measured conviction not a blind gamble. Berkshire does not have SP500 or NASDAQ as benchmarks, they don’t have benchmarks at all. They are deploying capital to beat the US treasury returns. 

4. Room to buy more if the bubble pops

The position is large but still a manageable share of Berkshire's liquidity. At $36.6 billion against $365.5 billion in cash,[2] it is roughly a tenth of the cash pile. If AI-related stocks fall, Berkshire has the means to add. It also keeps a firm floor. It will not repurchase its stock if that would push cash, equivalents and Treasury bills below $30 billion.[8] Even after a net equity-buying quarter and a $6.8 billion cash acquisition of Taylor Morrison, a very large reserve remains. Retail investors gambling their entire net worth or a substantial portion of it is not wise.

The takeaway

Berkshire's logic is not "AI is the future." It bought a cash-generating franchise at a size its balance sheet can easily carry, held for years, with plenty of capital left if prices fall. The risk is that heavy AI spending does not pay off. Berkshire seems to have sized the position with that risk in mind.

Sources

[1] Quartz, "Warren Buffett's Berkshire Hathaway buys Google stock for first time," https://qz.com/warren-buffett-berkshire-hathaway-buys-google-stock

[2] Forbes, "Berkshire Hathaway Turns Buyer As Buffett Backs Alphabet," Aug. 15, 2026, https://www.forbes.com/sites/bill_stone/2026/08/15/berkshire-hathaway-turns-buyer-as-buffett-backs-alphabet/

[3] CNBC, "Warren Buffett tells CNBC he initiated Berkshire Hathaway's investment in Alphabet," July 15, 2026, https://www.cnbc.com/2026/07/15/warren-buffett-tells-cnbc-he-initiated-berkshire-hathaways-investment-in-alphabet.html

[4] The Motley Fool, "Berkshire Hathaway Quietly Built a $37.4 Billion Stake in Alphabet," Oct. 14, 2026, https://www.fool.com/investing/2026/10/14/berkshire-quietly-built-374-billion-stake-in-apple/

[5] Barchart, "Berkshire Just Made a Huge Alphabet Bet," https://www.barchart.com/story/news/3930796/berkshire-just-made-a-huge-alphabet-bet-should-you-follow

[6] The Motley Fool, Aug. 2, 2026, https://www.fool.com/investing/2026/08/02/warren-buffett-regretted-not-investing-in-alphabet/

[7] The Motley Fool, Dec. 8, 2025, https://www.fool.com/investing/2025/12/08/5-stocks-warren-buffett-is-betting-big-on-for-2026

[8] Berkshire Hathaway Form 10-Q for the quarter ended June 30, 2026 (SEC EDGAR), https://www.sec.gov/Archives/edgar/data/0001067983/000119312526341032/brka-20260630.htm