3 Artificial Intelligence (AI) Stocks Warren Buffett Might Buy If He Were a Tech Investor

The article argues that Nvidia, Microsoft, and ASML are Buffett-style tech stocks with strong moats and exposure to AI. It highlights Nvidia’s trailing 12-month revenue of $216B and net income of $120B, Microsoft’s more than $305B in revenue with about a 40% margin, and ASML’s near-monopoly in EUV lithography with 16% sales growth last year. The piece is broadly constructive on these companies, but it is opinion/commentary rather than a new company-specific catalyst.

Analysis

The real signal here is not “Buffett likes tech” but that AI capex is still in an early industrial-cycle phase where the market is paying up for the toll collectors, not the application layer. NVDA and ASML are the cleanest beneficiaries because they sit closest to the physical bottlenecks; that gives them pricing power even if end-demand growth moderates. MSFT is more nuanced: it has the highest quality recurring cash flow, but AI monetization is likely to be slower and more visible in margin preservation than in near-term revenue acceleration.

Second-order effects favor the ecosystem around these names more than the headline stocks over a 1-3 month horizon. If hyperscalers keep spending, the constraint moves from “who has the best model” to “who can source enough advanced compute, lithography, and power infrastructure,” which supports the semi-capex chain and makes shortages more durable than consensus expects. That also creates a loser set: legacy chipmakers and software vendors without distribution or data moats will likely see the market assign them a higher AI adoption hurdle, not lower.

The main risk is a valuation air pocket, not a fundamental collapse. NVDA is the most exposed to any pause in capex guidance because expectations are already front-loaded; a single quarter of sequential growth deceleration could compress multiples sharply even if fundamentals remain strong. ASML has the best structural moat but the slowest catalyst realization, so it can underperform in momentum-driven tape despite being the most Buffett-like on a 2-5 year view.

Consensus may be underestimating how long AI spend can remain concentrated in infrastructure winners before moving to app-layer monetization. If that lag persists another 2-4 quarters, the right trade is not a broad AI basket but a barbell: own the tollbooths, avoid the high-beta “AI story” names with weak recurring revenue. Buffett framing is helpful, but the more important takeaway is that quality and scarcity are reasserting themselves inside tech, which usually supports relative performance even when absolute multiples compress.

Key Decisions for Investors

• Stay long NVDA into the next earnings cycle, but use call spreads rather than outright stock; the setup is favorable if capex guide remains firm, but valuation makes downside asymmetric on any sequential slowdown.
• Initiate a 6-12 month long ASML / short a basket of non-moat semiconductor equipment or AI-enabler laggards; ASML should outperform if the market keeps rewarding scarce supply chain bottlenecks.
• Own MSFT as the defensive AI compounder, but treat it as a lower-volatility quality anchor rather than a momentum trade; best entry is on post-print pullbacks where AI monetization skepticism compresses the multiple.
• Avoid chasing broad AI application names with weak distribution or thin margins; pair them short against NVDA or MSFT to express the view that infrastructure monetization will outrun software monetization over the next 2-4 quarters.
• If NVDA rallies hard into earnings, consider taking partial profits and rolling into longer-dated upside via calls; the risk/reward is best when implied volatility is elevated but secular demand remains intact.