Palantir's Co-Founder Just Dumped His Entire Nvidia Stake. Should Investors Follow Suit?

Nvidia reported blowout Q3 results with $57 billion in revenue versus an expected $54 billion, representing 62% year-over-year growth and accelerating from Q2, and CEO Jensen Huang signaled cloud GPUs are effectively sold out, implying supply constraints amid rising hyperscaler capex for 2026. Filings show Peter Thiel sold his Nvidia stake (previously ~537,000 shares) and a large Tesla position while redeploying capital into Microsoft and Apple, a move that created market attention but, per the piece, does not reflect weakening demand for Nvidia. The article argues Nvidia's forward P/E is comparable or cheaper than Microsoft and Apple, suggesting the company's fundamentals and near-term growth outlook remain robust despite insider selling.

Analysis

Market structure: Nvidia and its ecosystem (TSMC, ASML, memory vendors and cloud providers like MSFT/AMZN) are primary beneficiaries as Q3 showed 62% YoY revenue growth and CEO commentary that cloud GPUs are sold out — implying pricing power and allocation-based scarcity into 2026. Incumbent CPU/accelerator vendors (Intel, legacy on-premise OEMs) and discretionary-capex end markets (some EV spend e.g., TSLA exposure) are relatively hurt as compute capex rebalances toward hyperscalers and AI-specific silicon.

Risk assessment: Tail risks include export controls, a hyperscaler capex pullback, or foundry delays (TSMC capacity >90% util.), each capable of removing 20–50% of near-term upside. Time horizons: immediate (days) will be driven by headline volatility and IV; short-term (3–6 months) by capex guidance and inventory digestion; long-term (12–36 months) by fab capacity add/removes and competition from custom accelerators.

Trade implications: Primary actionable trades are concentration into NVDA and MSFT for exposure to cloud GPUs and application-layer capture, with hedges for convex downside. Use defined-risk option structures (3–9 month call spreads) to control IV exposure; consider pair trades: long NVDA/MSFT vs short TSLA or cyclic chip suppliers that fail to win GPU share to neutralize beta.

Contrarian angles: The market underestimates how quickly supply expansions (TSMC node adds, Nvidia A100/H100 successors) could normalise margins — compressing forward returns in 12–24 months. Conversely, Thiel’s selling is likely portfolio rebalance/liquidity, not a signal of AI demand failure; therefore oversold pullbacks in NVDA are tactical buy windows, not structural sell signals.

Key Decisions for Investors

• Establish a 1.5–2.0% portfolio long exposure to NVDA using a 3–6 month bull-call spread (buy ATM call, sell 20–30% OTM) to capture further revenue momentum while capping premium; reduce position by 50% if NVDA issues guidance cut or cloud backlog comments reverse within next 90 days.
• Allocate 1.5–2.5% to MSFT via equity or 6–12 month long-call (25% OTM) to play cloud GPU demand and software capture; hedge delta by shorting 0.5–1.0% TSLA equity or put spread to offset market beta given recent insider selling.
• Trim TSLA exposure by 30–50% within 10 trading days and redeploy 0.5–1.0% into PLTR (long equity) as a software/high-margin AI exposure; set stop-loss on PLTR at -25% and reassess after next 2 quarterly contract updates.
• Monitor three binary catalysts over 30–90 days (Nvidia backlog/capacity commentary, hyperscaler 2026 capex revisions, new export-control announcements). If any single catalyst changes negatively (e.g., hyperscaler capex guidance cut >15% yoy), cut NVDA exposure by at least 50% and shift to defensive large-cap tech (AAPL/MSFT) within 7 trading days.