Berkshire Hathaway Has 28% of Its Portfolio in These 3 Warren Buffett-Approved AI Stocks

Buffett and Greg Abel have concentrated Berkshire's AI-linked holdings to 28% of invested assets, with Apple at 21.4% and Alphabet at 6.8%, while Nucor remains a small 0.3% position. The article is constructive on Apple and Alphabet due to AI-related product and cloud growth, but more cautious on Nucor after Abel sold some shares ahead of strong earnings. Overall, this is primarily portfolio commentary rather than a fresh catalyst, so near-term market impact should be limited.

Analysis

The real signal here is not “Buffett owns AI,” but that Berkshire is quietly expressing a preference for AI monetization models with pricing power and capex discipline over pure infrastructure beta. Alphabet is the cleanest beneficiary because it monetizes AI twice: once through better ad relevance and again through cloud demand, which means the market can underestimate earnings leverage if margin expansion persists for another 2-3 quarters. Apple is more of a monetization optionality story than a direct AI winner; if upgraded AI features shorten replacement cycles even modestly, the multiple can stay elevated, but at this valuation the stock is already discounting a lot of that benefit.

The second-order effect most investors will miss is the supply chain squeeze from AI capex. Rising memory and component costs are a hidden tax on consumer hardware, which helps larger platforms with negotiating leverage and diversified revenue streams while pressuring smaller device makers and assemblers. That creates a relative winner/loser spread: high-quality hyperscalers can pass through cost inflation, while hardware names without services attach rates face margin compression.

Nucor is the most interesting contrarian line item because it is really a lagging beneficiary of data-center buildout, not an AI compounder. The setup is attractive only while construction demand stays ahead of new supply; once the new capacity comes online and pricing normalizes, earnings can mean-revert quickly. That makes the stock more of a tactical trade on backlog/pricing momentum than a durable AI structural long.

From a positioning perspective, the crowd is probably still underappreciating Alphabet relative to Apple. Apple is the consensus quality hold, but the forward multiple already leaves little room for execution slippage, whereas Alphabet’s valuation still looks compatible with sustained mid-teens earnings growth if cloud margins continue to expand. The main risk to the bullish thesis is not product failure but a demand fade in cloud or a pause in ad spend; that would hit both revenue and multiple simultaneously over a 6-12 month horizon.

Key Decisions for Investors

• Overweight GOOGL vs. AAPL on a 3-6 month horizon; pair trade long GOOGL / short AAPL to express the view that Alphabet has more operating leverage to AI monetization while Apple is already priced for perfection.
• Buy GOOGL on any 5-8% pullback, targeting a 12-18 month hold; upside comes from cloud margin expansion and AI-driven search monetization, with downside protected by diversified cash flows.
• Avoid adding to AAPL at current levels; if long, consider selling covered calls 1-2 quarters out to harvest premium against a high multiple and limited near-term catalyst density.
• Treat NUE as a tactical trade only: long for 1-2 quarters if data-center backlog and pricing remain firm, but use tight stops because new capacity could pressure margins into next year.
• For more defensive AI exposure, prefer BRK.B as an indirect allocator over single-name hardware; it reduces valuation risk while preserving exposure to the strongest secular AI beneficiaries.