AMD's Data Center Business Is Skyrocketing. The Stock Is Falling. Here's Why.

AMD’s Q2 data center revenue rose 107% Y/Y (from 57% last quarter) alongside improved margins to a 17% operating margin (up from a 2% loss a year ago), but the quarter still failed to “wow” the market. With AMD priced at ~70x forward earnings (~37x 2027 earnings), the sell-off is attributed more to valuation and expectations than business strength, especially versus Nvidia’s data center growth (92% Y/Y) and stronger sequential gains. The article also notes year-over-year comparisons were aided by the U.S. ban on chip sales to China, which hurt AMD and Nvidia last year.

Analysis

This is less a verdict on AMD’s operating trajectory than a reminder that the stock is still trading like a long-duration call option on flawless AI share gains. In that setup, good execution only supports the multiple; it does not expand it unless the company proves a step-function improvement in forward revenue or margin inflection over the next 2-3 quarters. The relative winner is NVDA, not because AMD is weak, but because the market is increasingly rewarding cash-earnings visibility and platform lock-in over catch-up growth.

The second-order effect is capital rotation inside semis: high-beta AI laggards can de-rate even on strong prints if they don’t outperform the highest bar in the group. That can pressure baskets like SMH/SOXX through factor de-risking, while reinforcing the “quality AI compounder” trade in NVDA and select infrastructure names. One nuance: if hyperscalers continue multi-sourcing to preserve negotiating leverage, AMD can still win sockets without ever regaining premium multiple status.

Near term, the key catalyst is not the last quarter but the next guide and the next two cloud capex read-throughs. Any evidence that AMD’s data-center growth is decelerating on a sequential basis, or that margins stall before the next product ramp, would justify further multiple compression. The contrarian miss is that AMD’s margin inflection is real enough to matter structurally; if operating margin holds in the mid-teens and keeps climbing, the current selloff could become the entry point rather than the warning sign.

Key Decisions for Investors

• Favor a relative-value pair: long NVDA / short AMD for 6-12 weeks, sized modestly, to express the thesis that earnings power is more durable in NVDA while AMD remains vulnerable to multiple compression; cut the pair if AMD’s next guide implies sustained >15% sequential DC growth or NVDA misses on q/q data-center momentum.
• On any AMD relief rally, consider buying 1-2 month put spreads rather than shorting outright; the risk is that the market has already over-discounted a reset and the next product-cycle update re-rates the stock back up.
• Do not add to broad semiconductor beta via SOXX/SMH until the next set of hyperscaler capex comments confirms AI demand is broadening; if the market starts de-risking high-multiple semis, the ETF can underperform even without an earnings deterioration.
• Set an alert for AMD margin and FY guide revision: if operating margin holds above the mid-teens and the company reiterates or raises forward data-center growth, the bear case weakens materially and the selloff becomes a buying opportunity rather than a fade.