I Own Nvidia, Microsoft, and Meta. Here's What I'm Doing With All 3 Right Now.

Stocks of Nvidia, Meta and Microsoft are down YTD through the week ending Mar. 20, 2026 (Nvidia -7%, Meta -10%, Microsoft -21%), reflecting cooling investor enthusiasm for AI despite strong company performance. Microsoft posted fiscal Q2 sales of $81.3B (+17% YoY), a P/E of 23, and disclosed $37.5B in capex (+66% YoY) to expand AI cloud capacity, which the author views as a buying opportunity and says he is adding to MSFT. Meta forecast 2026 capex of $115–$135B (vs $72.2B in 2025) and reported Q4 revenue of $59.9B (+24% YoY), while Nvidia delivered fiscal 2026 revenue of $215.9B (up from $130.5B) and launched the Vera Rubin GPU for inference, with CEO estimates of ~$1T in GPU orders by end-2027; the author recommends buy-and-hold for these AI leaders.

Analysis

Market skepticism on AI spending has created a two-speed outcome: capital-intensive owners of cloud scale (Microsoft, Meta) plus the dominant accelerator designer (Nvidia) structurally win, while commodity silicon and small middleware vendors face margin compression and demand bifurcation. The second-order beneficiaries are capacity enablers — hyperscaler ops teams, specialized datacenter integrators, and firms that lower power/cooling costs — which will see multi-year contract tailwinds even if headline multiple compression continues.

Key risks are asymmetric and calendarized: near-term (weeks–quarters) the story is volatility around earnings and capex guidance that can compress multiples quickly; medium-term (6–24 months) model-level innovations or price/performance breakthroughs from competitors could materially reduce incremental hardware demand. Regulatory, ad-monetization or macro-driven advertising slowdowns represent orthogonal catalysts that could impair Meta and, indirectly, the accelerator revenue cycle if advertising budgets retrench.

From a positioning standpoint, the market appears to be de-risking exposure to “all-AI” beta rather than the actual profitable sources of that AI (proprietary accelerators, long-term cloud commitments). That creates a window to express concentrated, asymmetric upside into semiconductor scarcity (NVDA), durable cloud cashflows (MSFT), and a targeted, lower-cost convex bet on Meta’s ad/engagement rebound — while shorting legacy silicon exposure (INTC) as a hedge against displacement and execution risk.

Key Decisions for Investors

• Buy NVDA 9–12 month LEAP calls (or a 12-month 1x long call position) financed by selling 1–3 month covered calls against ~25% of the notional. Timeframe: 6–12 months. R/R: target 30–60% upside; max downside = premium paid (size as 3–5% portfolio). This captures continued GPU scarcity and inference adoption while monetizing near-term IV.
• Pair trade — Long MSFT stock (or 9–12 month calls) / Short INTC stock sized 2:1 by notional. Timeframe: 6–18 months. R/R: asymmetric — MSFT upside via durable cloud contracts vs INTC downside from share loss and execution risk; set stop-loss at 12% adverse move in the spread.
• Tactical Meta trade — Buy a 12-month call spread on META (OTM buy / further OTM sell) sized small (1–2% portfolio). Timeframe: 6–12 months. R/R: limited-cost upside (target 25–40% on spread); use if ad metrics or ARPU inflection appears in next two prints.
• Systemic hedge — Purchase a 3–6 month NDAQ put spread or small S&P put spread sized 10% of gross long tech exposure. Timeframe: 0–3 months. R/R: protects against market-wide derisking that would erode option-funding and correlated tech positions while capping hedge cost.