AMD and Nvidia Split the AI Chip Market. 1 Stock Is Far Cheaper Today.

Nvidia's data-center revenue grew 117% in fiscal Q2 2027 to $89 billion, versus AMD's 107% growth to $6.7 billion in its Q2; the periods ended July 26 and June 30, respectively. The article says AMD is valued at roughly three times Nvidia on forward earnings despite Nvidia's faster growth, and argues Nvidia is the better long-term investment, while noting that any valuation convergence could come through an AMD decline or Nvidia rise.

Analysis

The relative-value case is plausible, but the stated valuation gap is not investable until the inputs are checked: forward P/E can diverge materially with estimate dates, normalized earnings, share-based compensation, and different segment mixes. Treat “three times” as an article claim, not a verified constant. Likewise, one quarter’s growth rates—on different fiscal calendars and business scopes—do not establish a durable earnings-growth advantage.

The key mechanism is expectation-setting. NVDA’s premium can remain supported if its software, networking, and full-system offering converts demand into durable pricing power; AMD could still outperform if it wins meaningful second-source deployments, even while growing more slowly in absolute dollars. Hyperscaler custom silicon is a real ceiling on accelerator demand and bargaining power, rather than a technology that can simply be assumed away. At the supply-chain level, constrained advanced packaging or high-bandwidth memory could limit shipments for both vendors and make customer allocation and delivery timing more important than nominal orders.

Near term (days), positioning and valuation claims may drive relative price action more than new fundamentals. Over 1–3 months, watch company guidance, independent customer spending signals, and revisions to data-center revenue and earnings estimates. Over 6–18 months, the test is whether NVDA sustains monetization and whether AMD converts design wins into shipped, profitable volume. The contrarian risk to the article’s thesis: a large valuation discount may already reflect a higher probability of share gains at AMD, while NVDA’s scale raises the bar for continued growth. Thesis weakens if NVDA’s estimates or margins are revised down faster than AMD’s, or if AMD’s deployments and estimates accelerate enough to justify its premium.

Key Decisions for Investors

• Watchlist, not an unconditional pair: verify current consensus earnings, estimate dates, and comparable forward multiples before positioning. If the asserted valuation gap persists on consistent inputs, consider a modest long NVDA / short AMD relative-value trade rather than an outright directional bet; cap risk because valuation convergence can take time or occur through NVDA underperformance.
• Use the next 1–3 months of guidance and estimate revisions as the trigger. Add only if NVDA’s data-center earnings outlook remains resilient and AMD does not show a material acceleration in profitable accelerator shipments; avoid treating announced design wins as realized revenue.
• Falsify the relative-value thesis if NVDA faces sustained downward estimate or margin revisions while AMD reports conversion of deployments into revenue and improving earnings expectations. Also reassess if hyperscaler custom-silicon investment meaningfully reduces GPU demand or pricing power.
• Monitor advanced-packaging and HBM availability alongside customer capex commentary: supply constraints can defer revenue rather than eliminate demand, but could also shift share between vendors. Do not infer vendor-specific exposure without shipment and allocation data.