Nvidia Trades at 21 Times Forward Earnings. Is the World's Biggest Artificial Intelligence (AI) Stock Actually a Value Play?

Nvidia now trades at ~21x forward earnings (down from >40x a few months ago) after record revenue of $215B and net income of $120B in the latest fiscal year, with gross margins >70% and management pointing to orders implying $1 trillion in revenue through 2027. The company continues to lead AI chip innovation (Blackwell, Blackwell Ultra, Vera Rubin system upcoming) and is positioned to capture inference and AI agent workloads, making it attractive to both growth and value investors despite recent stock weakness.

Analysis

The market is re-pricing AI exposure from a pure-trainings-capex narrative toward an inference-and-deployment story, which changes the unit economics across the stack: lower per-inference compute needs, higher aggregate inference transactions, and a longer lifetime for deployed accelerators. That combination favors companies with strong software/hardware integration and aftermarket services (provisioning, fleet management, inference orchestration) while eroding near-term ASP momentum for discrete training-focused silicon.

Expect two second-order supply dynamics over the next 6–24 months: (1) a growth in demand for medium-power edge and inference accelerators that will open TAM to alternative fabs and startups, compressing pricing at the high end; (2) continued bottlenecks in advanced packaging and HBM supply that will create episodic scarcity and margin moves for incumbents and their suppliers. These dynamics produce lumpy revenue cycles for dominant chip vendors even as underlying structural demand rises.

Key tail risks cluster around: accelerated competitive verticalization by hyperscalers (custom accelerators displacing merchant silicon), export-controls/regulatory segmentation that forces geographic product forks, and a rapid softness in enterprise CAPEX if cloud providers shift more to OPEX-heavy inference rental models. Timeframes matter: catalysts that can move price materially are quarterly guidance changes (days–weeks), supply/capex shifts (months), and architectural share shifts from hyperscaler vertical designs (12–36 months).

From a positioning standpoint, the best opportunities are asymmetric option structures and relative-value pairs that isolate share gains versus cyclical inventory and margin moves. Avoid unilateral carry trades that assume linear re-acceleration; volatility is the vehicle where conviction translates to high expected payoff with defined downside.

Key Decisions for Investors

• Long-dated asymmetric NVDA LEAPs: Buy NVDA Jan-2028 calls ~25–30% OTM (size 2–4% of equity portfolio) financed by selling NVDA Sep-2026 calls (1:1 or 2:1 financing). Rationale: captures multi-year share capture in inference and software monetization while financing time decay from near-term uncertainty. Risk/reward: limited premium spent vs. >3x upside if NVDA retains pricing power; downside is premium loss if adoption slows in 12 months.
• Relative-value pair: Long NVDA / Short INTC (equal-dollar) for 6–18 months. Rationale: isolates exposure to merchant-accelerator share wins while hedging macro and semiconductor cyclicality. Risk/reward: target a 25–40% relative return if NVIDIA extends lead; stop-loss if NVDA underperforms INTC by 10% on a relative basis (signals competitive surprise).
• Premium-collection hedge: Sell cash-secured NVDA 6-month puts 8–12% OTM sized to buy-on-assignment. Rationale: collect elevated premium while setting an attractive effective entry if a pullback materializes around quarterly guide risks. Risk/reward: collects premium (buffer ~premium%) but risks full equity assignment if NVDA gaps lower beyond strike; treat as disciplined buy-on-assignment strategy.
• Customer-side play: Buy AMZN 9–12 month calls (modest size). Rationale: asynchronous capex patterns among hyperscalers mean cloud providers convert backend CAPEX into durable revenue for themselves; AWS capture of inference spend creates outsized cashflow optionality versus on-prem exposure. Risk/reward: benefits if inference monetization accelerates; downside is cloud capex softness or margin compression at AMZN.