Angry Nvidia CEO Jensen Huang says: Stop calling all my investments and acquisitions ...

Nvidia has committed a further $2 billion to cloud provider CoreWeave to fund land, power and infrastructure for AI ‘factories,’ reinforcing a strategic partnership in which CoreWeave rents Nvidia GPUs rather than Nvidia building its own cloud. CEO Jensen Huang defended the moves against claims of circular financing, noting Nvidia’s stakes (including a prior $250 million IPO anchor at $40/share) are a small fraction of the capital required for AI expansion; CoreWeave raised $1.5 billion in its IPO and Nvidia’s chips will underpin the buildout. Huang also highlighted the scale of industry investment plans, citing OpenAI’s projected $1.4 trillion spend over eight years, underscoring continued heavy upstream demand for Nvidia hardware.

Analysis

Market structure: Nvidia (NVDA) and specialized cloud providers like CoreWeave (CRWV) are clear winners as strategic investments lock demand for Nvidia GPUs and raise switching costs for competitors; expect Nvidia ASPs and channel leverage to remain elevated, supporting gross-margin resilience over the next 12–24 months. Traditional hyperscalers may see margin pressure or be forced to pre-commit capex, benefitting data‑center power/utility names and equipment suppliers, while GPU rivals (AMD, INTC) face incremental pricing pressure and potential share loss.

Risk assessment: Key tail risks are regulatory/antitrust action against vertical partnerships, tighter US export controls, and a CoreWeave concentration shock (e.g., >30% revenue tied to one supplier), any of which could knock 20–40% off implied growth expectations. Timing: expect immediate stock volatility around disclosures/earnings (days–weeks), funding and build cycles to play out over months, and structural market-share shifts over 1–3 years; watch capital-raise cadence and energy-grid constraints as hidden dependencies.

Trade implications: Direct play is a tactical long NVDA (size 2–4% portfolio) on a 6–12 month horizon, using staggered buys on 5–10% pullbacks and capped-cost call spreads (buy 9–12m calls 15–25% OTM, sell 50–80% OTM) to limit spend. Pair trade: go long NVDA and short AMD (equal notional, 0.75% each) for 3–9 months to capture relative GPU pricing power; add a 0.5–1% portfolio allocation to CRWV for 12–24 months, trim if CRWV rallies >40% or reports customer concentration >30%.

Contrarian angles: Consensus understates regulatory and concentration risk — circular financing may invite scrutiny that compresses multiples by 15–30% if labeled anti‑competitive. Historical parallels to past hardware-cloud entanglements (early 2000s) show fast initial wins followed by regulatory/contract renegotiation; if NVDA’s investment-to-revenue ratio rises materially (>5% of partner market cap invested), be prepared to rapid de-risk positions.

Key Decisions for Investors

• Establish a 2–4% net long position in NVDA for a 6–12 month horizon; dollar-cost into any 5–10% pullback, and hedge with 0.5–1% portfolio cost using 9–12 month call spreads: buy calls 15–25% OTM and sell calls 50–80% OTM to cap premium outlay.
• Add a 0.5–1% conviction long in CRWV for 12–24 months to play AI factory demand; set an automatic trim at +40% price appreciation or if CRWV discloses >30% revenue dependency on a single customer.
• Implement a 0.75% long NVDA / 0.75% short AMD pair trade (equal notional) for 3–9 months to capture relative GPU pricing dynamics; cover/flip if AMD outperforms NVDA by 15% or if NVDA reports material regulatory inquiries.
• Allocate 0.5–1% portfolio to tail hedges: buy 3–6 month 10–15 delta puts on NVDA (or equivalent index options) to protect against a 20–30% downside event tied to regulatory/export control shocks; if DOJ/FTC opens formal inquiry within 60 days, reduce NVDA exposure to <=1%.