Buffett’s Berkshire quietly unloads AMZN just as Al spending soars | Official Website of Louis Velazquez, entrepreneur, finance guy and tech innovator

If you just look at the headlines about AI, you’d never expect Berkshire Hathaway to be backing away from Amazon right now.

In Warren Buffett’s final quarter as CEO, Berkshire’s latest Form 13F for the period ended Dec. 31, 2025, shows a massive cut to its Amazon.com stake, reducing the position by more than 77% in share terms.

The 13F‑HR filed with the Securities and Exchange Commission on Feb. 13, 2026, confirms that Berkshire reported 110 line‑item equity positions with an aggregate fair value of about $274.2 billion, setting the official record of its public‑stock portfolio at year‑end.

According to a breakdown of the filing cited by both Investopedia and third‑party portfolio trackers, Berkshire sold the bulk of its Amazon holdings in Q4 2025, turning what had been a high‑profile bet on e‑commerce and cloud into a much smaller residual stake.

That’s not the move you’d expect when Amazon Web Services is powering a large share of the AI boom.

Berkshire sold 77% of its Amazon stake in Q4 2025.

Photo by Nathan Stirk on Getty Images

AI spending explodes while Berkshire steps aside

The timing of this Amazon sale makes the story more interesting. You have AI capital spending surging almost everywhere you look, and Berkshire is shrinking exposure to one of the biggest corporate spenders.

Alphabet, Microsoft, Meta, and Amazon are expected to spend nearly $700 billion combined this year to fuel their AI build‑outs, highlighting what it called an “AI cash bonfire” that is testing investors’ patience, CNBC reported.

Wall Street also expects those four to spend between $635 billion and $665 billion on AI and related capex in 2026, Yahoo Finance noted.

Global figures tell the same story. Worldwide AI spending is projected to reach about $2.52 trillion in 2026, according to Gartner. This is a 44% jump from the prior year, with AI infrastructure alone adding roughly $401 billion to that total as companies build out AI‑optimized servers and data centers.

Amazon is in the thick of that, planning roughly $200 billion in capital expenditures in 2026, much of it tied to AI and cloud infrastructure at Amazon Web Services, according to Intellectia.ai’s summary of Amazon’s latest guidance.

So Berkshire is stepping back just as the numbers go from large to almost hard to process. To me, that looks intentional, not accidental.

What the 13F reveals about Berkshire’s AI stance

The SEC filing is just a snapshot, but if you line up Berkshire’s recent 13Fs, a pattern jumps out. The firm is not fleeing AI altogether. It’s reshaping how it wants to get paid for that risk.

Apple remains Berkshire’s largest single equity holding, even after additional selling in 2025, with more than 227 million shares still on the books at year‑end, according to 13F data compiled by Kavout. That’s decades of cash‑flow visibility, plus Apple’s own AI plans in devices and services, without the same direct data‑center capex burden the cloud giants carry.

The same review, along with coverage from Investopedia and others, shows that Berkshire cut Amazon sharply while making room for a new multi‑billion‑dollar position in Alphabet.

More AI Stocks:

Berkshire built a roughly $4.3 billion stake in Alphabet, vaulting the company straight into Berkshire’s top holdings by value, CNBC reported. That move came “two decades after” Buffett helped inspire Google’s IPO but chose not to buy in, highlighting how selective Berkshire is about tech exposure.

The Alphabet purchase is “an endorsement of the tech giant’s artificial intelligence efforts,” with CFRA analyst Angel Zino saying the deal “validates Google’s strong fundamentals and provides Berkshire exposure to a leading AI provider through Google Cloud and Gemini expansion,” Fox Business reported.

Alphabet is “emerging as a leader in AI, showcasing enhancements in Google Search and the Gemini chatbot,” and its valuation remains acceptable despite a huge rally, Yahoo Finance argued.

So Berkshire is not allergic to AI. It’s choosing where it wants AI risk to live, and on what terms. That’s a subtle but crucial difference for anyone building their own portfolio.

Berkshire’s move: Why Amazon, and why now?

From the outside, it’s tempting to say Berkshire is walking away from Amazon at precisely the wrong time. I don’t buy that.

Berkshire’s Q4 2025 13F shows that this was not a one‑off sale, with the firm trimming Amazon for several quarters before taking a much bigger bite out of the position at year‑end, Investopedia reported. That pattern looks like a controlled, multi‑step de‑risking process, not a sudden change of heart.

On the business side, AWS is still growing fast. AWS posted about 24% year‑over‑year revenue growth in the most recent quarter, pushing its annualized revenue run‑rate above $140 billion, Intellectia noted. Those are not the numbers of a company losing the AI race. They are the numbers of a business in full investment mode.

The question is not whether AI will matter for AWS. It’s whether Amazon can keep spending at a roughly $200 billion annual capex rate while still delivering the kind of compounding cash returns Berkshire prefers.

If your entire philosophy is to avoid situations where everything has to go right to make money, that kind of setup is a warning sign. Berkshire appears to be treating it that way.

How you can apply Berkshire’s playbook

When I look at these filings as a retail investor, the lesson is not “sell Amazon because Berkshire did.” The lesson is to separate the AI story from the stock story.

Here are a few practical ways to use that mindset.

Focus on how you get AI exposure, not how many AI stocks you own. Berkshire is keeping Apple, adding Alphabet, and shrinking Amazon, which shows plenty of AI upside can come through platform and ecosystem names, not just direct cloud spenders.
Track capex discipline as closely as you track AI announcements. Amazon’s planned $200 billion in 2026 capex and similar surges across peers look impressive on slides, but they only work for you if incremental returns exceed the cost of that capital.
Manage your own concentration risk. Berkshire’s steady trimming of Apple, Amazon, and big banks while boosting Alphabet and stockpiling cash shows a clear decision to prevent any one winner from dominating the portfolio.

Berkshire has been a net seller of stocks for multiple quarters, building a cash pile rather than leaning harder into a market driven by AI, megacap tech, and narrow leadership, TheStreet’s coverage noted. That is another way of saying, “We don’t have to own everything right now.”

If you are nervous about missing out on AI, I’d challenge you to think more like that.

Decide where you want your AI risk to live: in the chips that power models, in the cloud platforms that host them, or in the consumer and enterprise applications that turn AI into recurring revenue. Then size those positions so that one misstep doesn’t derail your whole plan.

The bigger signal behind Berkshire’s AI timing

This Amazon sale landed in the same quarter that marked the final 13F of the Buffett era. That matters symbolically, but I think it matters more as a roadmap for how Berkshire wants to face an AI‑heavy market.

Gartner’s projections of multi‑trillion‑dollar AI spending show just how big this wave can be. Yet the 13F Berkshire filed on Feb. 13, 2026, shows a firm willing to reduce its stake in a headline AI winner at the very moment the spending numbers turn spectacular.

To me, that’s a reminder that you can respect the power of a technology cycle without surrendering your discipline on valuation, capital intensity, and portfolio balance.

You don’t have to chase every name that benefits from AI. You just have to be deliberate about the ones you choose to own, and honest about the price you’re paying for that growth.