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

Berkshire Hathaway’s AI-linked holdings are concentrated in Apple (21.4% of invested assets), Alphabet (6.8%), and Nucor (0.3%), with Greg Abel reportedly tripling down on Alphabet and treating Apple as a core hold. The article highlights strong fundamentals for Apple and Alphabet, including Apple’s AI-enabled Siri refresh and Alphabet’s 63% Google Cloud revenue growth and margin expansion to 32.9%. Nucor remains a smaller, more cyclical AI/data-center beneficiary, though Abel has recently sold some shares.

Analysis

The hidden signal is not Buffett’s taste for AI; it’s Berkshire’s pivot toward picking the infrastructure and monetization layers of AI rather than the model layer itself. That favors firms with pricing power, distribution, or embedded enterprise demand, while leaving pure compute winners exposed to capex cycles and compression once supply catches up. In practice, the market is still paying up for the “AI narrative” at the chip/cloud layer, but the better risk-adjusted exposure here is to businesses that can turn AI into higher engagement, lower churn, or share gains without needing explosive unit growth.

Alphabet looks like the best asymmetry in the set because the market is still valuing it like a mature ad platform, not a beneficiary of AI-driven search monetization and cloud operating leverage. If cloud margins keep expanding and AI improves ad targeting/engagement, earnings revisions can compound for several quarters even without multiple expansion. The main risk is not competition from OpenAI-style apps; it’s that AI answer surfaces cannibalize query volume faster than Alphabet can reprice inventory, which would show up first as traffic mix deterioration before headline revenue weakness.

Apple is more of a hold than a fresh buy: the catalyst is a device upgrade cycle tied to AI features, but the stock already discounts a lot of that optionality. Margin pressure from component inflation matters because it can delay the exact gross margin expansion investors want from AI-enabled services. The second-order winner could be the App Store ecosystem and premium accessory suppliers, while the loser is any smartphone OEM trying to compete on AI features without Apple’s installed base and pricing power.

Nucor is the contrarian exposure: it benefits from AI capex through data center steel demand, but this is a cyclical winner masquerading as an AI stock. The market will likely underestimate how quickly its earnings can mean-revert once hyperscaler build rates normalize, especially if new capacity comes online into a softer pricing environment. That makes the name attractive tactically on dips, but poor as a long-duration AI compounder.

Key Decisions for Investors

• Long GOOGL vs short AAPL on a 3-6 month horizon: Alphabet has cleaner upside from AI monetization and cloud operating leverage, while Apple already trades with rich expectations and more execution risk; target a relative return, not outright beta.
• Add to GOOGL on any 5-8% pullback or post-earnings volatility: the setup is favorable for estimate revisions over the next 2-3 quarters if cloud margin expansion persists; use tight risk limits if search engagement metrics soften.
• Hold AAPL but do not chase: treat as a core-quality compounder, not an AI upside trade; if you need upside optionality, prefer call spreads only into product-event windows, with willingness to trim if forward P/E stays elevated above the mid-30s.
• Tactically buy NUE only on weakness and as a 6-12 month trade: the data-center demand story can support near-term earnings, but cyclicality argues for taking profits before new capacity starts to pressure margins; avoid adding if pricing peaks before 2027.
• Use BRK.B as a low-volatility expression of the basket, but recognize it is now more a capital allocation overlay than a direct AI trade; pair it against more expensive tech beta if you want reduced drawdown with some AI exposure.