Nvidia CEO denies ‘nonsense’ report he’s unhappy with OpenAI. ‘I really love working with Sam’

Nvidia CEO Jensen Huang said reports of a binding $100 billion investment in OpenAI are false, but confirmed Nvidia will make a very large, likely record, investment and remains supportive of OpenAI and CEO Sam Altman. Nvidia signed a letter of intent in September to potentially invest up to $100 billion for data centers and infrastructure, though Huang said that figure was never a commitment; OpenAI is reportedly targeting up to $100 billion in fundraising that could value the company above $800 billion and may consider an IPO in 2026. The developments—alongside Nvidia’s additional $2 billion deployment into CoreWeave—underscore continued heavy capital deployment across AI infrastructure and potential circular spending among chipmakers and hyperscalers, with material implications for capital allocation in the sector.

Analysis

Market structure: Nvidia (NVDA) is the clear winner — a material capital injection and sustained GPU demand preserve its pricing power and margin leadership, while OpenAI gains cheaper, secured capacity and deeper product-market fit. Hyperscalers (AMZN, GOOGL) and GPU cloud providers (CoreWeave/CRWV) benefit from increased model scale but face higher input costs and potential circularity in reported returns as capex flows between partners inflate revenue figures. Expect persistent positive demand shock for high-end GPUs that keeps spot allocation tight for 12–24 months and supports semi-capacity premiums of 10–30% versus pre-AI cycle levels.

Risk assessment: Tail risks include antitrust/export-control scrutiny of large inter-company investments, a regulatory clampdown if circular funding masks returns, or a rapid shift to custom silicon that reduces Nvidia TAM within 2–5 years. Immediate (days) risk is headline-driven volatility; short-term (weeks–months) risks center on fundraising announcements and Nvidia earnings; long-term (quarters–years) risks are competitive displacement and power/real-estate constraints increasing marginal costs. Hidden dependency: OpenAI’s growth is power/water and real-estate constrained — not just GPUs — creating non-linear marginal costs and capex dilution.

Trade implications: Tactical: overweight NVDA and cloud scalers (AMZN, GOOGL) to capture GPU-driven revenue; employ time-limited option structures to limit drawdown. Use 6–9 month call spreads on NVDA to capture outsized upside while selling nearer-term premium to finance cost; add selective long exposure to AWS/GCP (0.5–1% each) as a play on commercialized AI services. Rotate into utilities, power infrastructure suppliers, and data-center REITs over 3–12 months; trim cyclical ad-exposed tech names if capex diverts digital ad budgets.

Contrarian angles: Consensus underestimates regulatory and circular-spend risk — a large disclosed NVDA equity stake (>~$10–20B) could trigger deeper scrutiny and erode multiple expansion. The market may be overpricing perpetual linear scaling; historical analogues (Microsoft–OpenAI 2019) show high upside but long ROI tails — expect >18–36 months to monetization parity. If OpenAI’s pre-money exceeds ~$500–800B, odds of deal dilution, governance complexity, and a valuation reset rise materially.

Key Decisions for Investors

• Establish a 1–2% portfolio long position in NVDA via a 6–9 month call debit spread (buy ATM call, sell ~30% OTM call) to capture 30–60% upside potential; set a tactical stop-loss to cut exposure if NVDA falls >25% from entry within 3 months.
• Add 0.5–0.75% long positions in AMZN (AWS exposure) and 0.5% in GOOGL (GCP) via shares or 12‑month LEAPS; reduce to neutral if combined cloud GPU revenue growth misses consensus by >200 bps over two consecutive quarters.
• Purchase a 1% notional portfolio hedge: NVDA 6–9 month puts ~15–25% OTM (size to cap loss to ~1% portfolio) to protect against headline/regulatory blow-ups around the OpenAI financing in the next 3 months.
• Underweight/avoid direct exposure to small-cap AI infra names and private cloud GPU plays (e.g., CRWV-type exposures) — consider shorting a concentrated basket of negative‑FCF AI infra stocks equal to 0.5–1% notional if they rally >20% on partnership headlines; reassess after fundraising details are public.