Why Berkshire Buying More Alphabet Is a Value Trade, Not an AI Bet

As reported by Reuters and confirmed in Berkshire Hathaway’s Q2 2026 13F filing. The analysis below is our own.

Berkshire Hathaway spent the second quarter buying more Alphabet — roughly 24.5 million additional shares, lifting the position about 45% to nearly $28.2 billion. That makes Google’s parent one of Berkshire’s five largest equity holdings, sitting at 9.4% of the book alongside Apple, American Express, Coca-Cola and Bank of America. Those five names are now 66% of the entire portfolio.

Berkshire Hathaway Q2 2026 top-five holdings, Alphabet highlighted

The easy headline is “Buffett buys Google.” The more useful observation is stranger: the most famously technology-averse pool of capital on Earth — the fund that sat out the entire cloud era and only bought Apple once it behaved like a consumer-staples brand — has quietly made a Big Tech AI platform a top holding. And it did so, per the original disclosure earlier this year, partly to help fund Alphabet’s AI development. Value investing did not blink at AI. It underwrote it.

Why this is a value trade, not an AI bet

The two look identical from the outside and are opposite in logic. An AI bet pays for a future that may or may not arrive — you buy the story and hope the cash flow follows. A value trade buys cash flow that already exists and treats the AI upside as a free option the market has mispriced. Berkshire is doing the second. Alphabet throws off tens of billions in operating cash from Search and a structurally profitable ads engine today, at a multiple far below the pure-play AI names. You are not paying for Gemini. You are paying for the ads business, and Gemini comes in the box.

That is the whole trick of buying an incumbent into a technology shift. The downside is capped by a cash machine that exists whether or not the new thing works; the upside is a call option on the new thing, bought at no premium. It is the inverse of how most of the market is currently pricing AI — paying rich multiples for optionality and hoping the economics show up later.

What makes Alphabet the one incumbent that qualifies

Not every incumbent is a value-safe way to own AI. Alphabet is rare because it captures value at every layer of the stack at once, and each layer funds the next. It designs its own compute (TPUs), trains and owns frontier models (Gemini), and monetizes both through products that already print cash (Search, Cloud, YouTube). Most companies touch one layer and rent the rest. Alphabet owns the column — which means the AI build-out shows up on its own income statement as capability, not just as a bill from a supplier.

That is the structural reason a conservative investor can hold it without holding a story: its position in the stack converts the AI spend into a moat rather than a cost. The same capex that is a pure expense for a single-layer company is, for a full-stack owner, an investment in the thing that defends the cash flow underneath it.

The Enterprise AI Coordinate System
Free book · The Business Engineer
The Enterprise AI Coordinate System
Which layer of the stack captures the value — and which just pays the bill. The 65-page buyer’s map, free.
Get the free map →

The Abel signal underneath it

This was also the first full 13F under Greg Abel, and the quarter carried roughly $19.8 billion of net buying — a decisive move for a firm that had spent years letting cash pile up. Reading the Alphabet build inside that context, the signal is not “Berkshire discovered AI.” It is that the new regime is willing to define its circle of competence to include a technology platform, as long as that platform is bought on value terms rather than narrative ones. The discipline didn’t change. The map of what counts as a durable business did.

The lesson for everyone who is not Berkshire is the cleaner one: the winning way to own this cycle may not be the pure-play that is the story, but the incumbent whose existing cash flow lets you own the story for free. That only works when the incumbent occupies the right position in the stack — which is exactly the thing the market keeps failing to price.


Sources: Reuters · Berkshire Hathaway Q2 2026 13F (via Seeking Alpha, CNBC portfolio tracker) · Benzinga, Kingswell (holdings breakdown). Rankings from the 13F place Alphabet at 9.4% (fourth by single-class value; some outlets rank it third when combining both Alphabet share classes). The initial ~$10B stake was disclosed earlier in 2026.

Scroll to Top

Discover more from FourWeekMBA

Subscribe now to keep reading and get access to the full archive.

Continue reading

FourWeekMBA