AMD and Nvidia Just Experienced an Unprecedented Valuation Split

The article argues Nvidia offers a materially better valuation than AMD despite superior profitability and AI data-center exposure. Nvidia generated $89 billion of its $96.2 billion Q2 revenue from data centers, versus AMD's $6.7 billion of $11.5 billion; AMD's lower-margin Client & Gaming business represented $3.8 billion of revenue with a 15% operating margin. The author contends AMD trades at more than twice Nvidia's valuation while Nvidia has stronger margins, faster growth, and a more favorable business mix, favoring Nvidia shares over AMD.

Analysis

The relevant trade is not a simple quality comparison; it is a debate over whether AMD can convert AI accelerator revenue into a durable gross-margin and operating-margin step-up. AMD retains material exposure to cyclical PC, gaming, and embedded end-markets, so a softer consumer or industrial cycle can dilute consolidated earnings even if MI-series accelerator sales grow rapidly. NVDA's valuation support is more directly tied to hyperscaler capex, networking attach, and the durability of its software/ecosystem moat; this creates cleaner earnings sensitivity but also greater downside if cloud customers signal digestion.

A long NVDA/short AMD pair is attractive only if the current forward-EPS and enterprise-value multiples preserve a meaningful spread after adjusting for growth. The article's valuation framing is not independently sufficient: consensus estimates may already embed a sharp AMD AI ramp and a deceleration in NVDA growth, making the pair vulnerable to an AMD design-win or supply-availability surprise. Over the next 1-3 months, hyperscaler capex commentary, AMD accelerator shipment/MI300-family margin disclosures, and NVDA rack-scale product availability are the key catalysts; over 6-18 months, the decisive variable is whether AMD wins recurring inference deployments rather than isolated training qualifications.

Consensus may underweight NVDA's own transition risk. A product-cycle shift toward more integrated rack systems can increase revenue per deployment but also raises execution complexity, working-capital needs, and customer qualification timing; any delayed systems ramp could temporarily compress the multiple despite intact long-term demand. Conversely, AMD need not reach NVDA-like margins for its shares to outperform: a credible path to sustained data-center mix expansion plus improving client demand could drive estimate revisions faster than a relative-valuation thesis anticipates.

Key Decisions for Investors

• Establish a modest 3-6 month long NVDA / short AMD dollar-neutral pair only after confirming the forward EV/EBIT spread remains above its 12-month median; target 10-15% relative return, with a 7% adverse relative-price stop.
• Express NVDA upside through a 6-month call spread rather than unhedged equity if implied volatility is below its one-year median; use a 10-15% out-of-the-money long strike and cap upside near a 25-30% move to limit premium outlay.
• Do not add to an AMD short ahead of earnings without verifying accelerator revenue, gross-margin bridge, and customer concentration data. Cover the relative short if AMD guides data-center revenue growth materially above consensus or demonstrates sustained accelerator gross margin accretion.
• Monitor MSFT, AMZN, GOOGL and META capex guidance as the highest-frequency read-through. Reduce NVDA exposure if two or more hyperscalers guide sequential AI infrastructure spending moderation or cite deployment bottlenecks.
• Treat any sharp AMD selloff as a watch item rather than an automatic short add: evidence of PC recovery or a major inference design win could create a favorable earnings-revision setup despite structurally lower margins.