Bill Gates' Foundation Holds Berkshire Hathaway as Its Top Stock, a Signal of Its Preference for Steady Compounders Over Flashy Tech

The Gates Foundation Trust reported a concentrated $34.4B equity portfolio (24 stocks) as of June 30, notably emphasizing industrials/consumer names over tech. Its Caterpillar position highlights the AI spillover into infrastructure: CAT is up 44% YTD, with Q2 revenue up 24% YoY to $20.5B, and it raised its dividend 8% to $1.63/quarter (0.8% yield). The article argues investors can capture parts of the AI boom via non-tech holdings, even as it notes tech’s higher volatility and the foundation’s underweighting of the sector.

Analysis

The only actionable signal here is not “buy what Gates owns,” but that capital is still flowing into businesses with durable pricing power, visible cash returns, and less dependence on terminal multiple expansion. That matters because in a regime where AI winners are increasingly crowded, the market is still underpricing the infrastructure layer: CAT is effectively a toll collector on data-center buildouts, and that revenue stream can surprise to the upside even if software/semis go sideways.

Second-order, the bigger loser may be the broad industrial cohort that does not participate in AI-related buildout intensity. If data-center capex stays concentrated, CAT should keep taking share in investor attention from slower-moving peers like DE, while contractors and rental fleets become the nearer-term beneficiaries of volume, not just OEMs. By contrast, WM and WMT are less about upside and more about capital preservation; they can outperform on a risk-off tape, but they are unlikely to re-rate without a clearer slowdown in growth-sensitive assets.

The contrarian miss is that 13F mimicry is stale by construction: the foundation’s holdings tell you what worked over the last quarter, not what compounds next. The more interesting question is whether the market is extrapolating CAT’s AI link too far into a full-cycle earnings stream. If hyperscaler capex or construction orders soften over the next 1-2 quarters, CAT can de-rate quickly even if fundamentals remain fine; the falsifier is continued backlog/order strength into the next print, especially if pricing holds and margins do not mean-revert.

Key Decisions for Investors

• Long CAT / short DE for 1-3 months as a relative-value expression of AI infrastructure exposure versus ag-cycle sensitivity; thesis breaks if Deere’s order book re-accelerates or CAT’s data-center mix slows.
• Do not buy BRK.B solely on the 13F; treat it as a stale ownership signal. Wait for a drawdown or operating catalyst rather than paying up for a passive-quality compounder.
• Use WM and WMT as defensive quality longs if the market broadens into risk-off; this is a ballast trade, not a catalyst trade, with better downside capture than chasing AI-related beta.
• If already long CAT after the run, consider a 3-6 month put spread or covered-call overwrite to harvest elevated expectations; invalidated if next quarterly orders/backlog and pricing both re-accelerate.