AMD Is Up Nearly 300% Over the Past Year. Is It Still a Buy?

AMD’s Data Center division grew 57% in Q1 and management projected a 60% CAGR over the next 3-5 years, but the stock now trades at 63x forward earnings. The article argues that despite strong AI-linked growth, AMD looks expensive relative to Nvidia, which is growing revenue 85% year over year and trades at about 24x forward earnings. The main message is valuation caution rather than an operating miss.

Analysis

The key market issue is not whether AMD can keep growing; it’s whether the market has already capitalized too much of that growth into the multiple. At ~63x forward earnings, AMD now needs near-flawless execution for multiple quarters, and that creates a setup where any deceleration in data center bookings or margin mix can hit the stock harder than the underlying business fundamentals would suggest. In other words, the stock has transitioned from a “show me” story to a “don’t miss” story, which is a much less forgiving regime.

The second-order winner is NVDA, but not just because it is cheaper on the stated forward multiple. The more important dynamic is that capital in AI semis tends to migrate toward the name with the most visible near-term monetization and strongest balance-sheet optionality, which makes NVDA the cleaner expression for investors wanting AI exposure without paying a premium for execution lag. If AMD’s multiple stays elevated, it also increases the odds that relative-return investors rotate from AMD into suppliers, infrastructure, or even software beneficiaries with less direct hardware cyclicality.

The contrarian angle is that the market may be underpricing duration: AI capex cycles often extend longer than consensus expects, and AMD’s data center ramp can keep surprising to the upside for several quarters even if growth normalizes later. The risk to the short-premium thesis is that multiple compression usually needs a catalyst — a guidance cut, margin miss, or evidence that hyperscaler demand is pausing — and absent that, overvalued stocks can stay expensive for months. So this is less a “short now” setup than a “own the cheaper winner, fade the crowded premium” setup.

For the next 1-2 quarters, the biggest reversal risk is that AMD’s growth trajectory remains intact while NVDA’s growth rate moderates faster than expected, which would keep the relative-value trade noisy. But if AI spending broadens beyond a small set of hyperscalers, AMD’s consumer exposure becomes less important and the stock could keep levitating despite the valuation. That makes timing critical: the bearish case is strongest on strength into earnings or after a sharp post-print run, not on weak tape alone.

Key Decisions for Investors

• Rotate from AMD into NVDA on strength: long NVDA / short AMD pair trade for 3-6 months, targeting multiple convergence if AMD de-rates from a premium growth multiple; stop if AMD continues to outperform on data center guide raises.
• If already long AMD, trim 25-50% into post-earnings rallies and redeploy into NVDA or a basket of AI infrastructure names; the risk/reward is now asymmetric against adding AMD at a premium valuation.
• For directional exposure, buy NVDA call spreads 2-4 months out rather than outright AMD calls; this keeps upside to AI demand intact while reducing exposure to valuation compression.
• Use AMD put spreads after any guidance-led gap up: 10-15% OTM, 1-2 quarter horizon, designed to monetize a reset in expectations rather than a catastrophic fundamental break.
• Monitor hyperscaler capex commentary over the next earnings cycle; if spending slows even modestly, reduce AMD exposure first, as the stock has the highest valuation sensitivity to any change in AI digestion pace.