AMD Shares Slide Despite Strong Growth. Is It Time to Buy the Stock on the Dip?

Advanced Micro Devices reported Q4 revenue of $10.27 billion, up 34% year‑over‑year, with data‑center revenue rising 39% to $5.4 billion and client & gaming revenue up 37% to $3.9 billion (client $3.1B, gaming $843M). Gross margin expanded to 54% (+300 bps) aided by a reversal of an MI308 China write‑down, adjusted EPS was $1.53 versus a $1.32 consensus, and the company guided Q1 revenue of $9.8 billion ± $300M (up ~32% YoY). The quarter was helped by $390 million of GPU sales into China that management described as "dynamic" and may not repeat (only $100M modeled for Q1), while AMD flagged a meaningful decline in semi‑custom gaming revenue in 2026 even as it gains share in CPUs and expands GPU deployments with eight of the top 10 AI firms and planned GPU deliveries to OpenAI in H2 2026.

Analysis

Market structure: AMD’s beat highlights two structural winners — AI GPU suppliers (AMD, NVDA) and hyperscalers/cloud (MSFT/GOOG/OpenAI customers) — while semi-custom gaming suppliers and console-dependent OEMs look vulnerable as AMD warns semi-custom will decline in 2026. The $390M China boost in Q4 is an irregular tailwind; absent repeatability, near-term revenue growth could undershoot current sentiment even as data‑center CPU/GPU demand remains structurally tight. Cross-asset: continued AI capex supports risk‑on equities, flattens yields on growth optimism, lifts implied vols for AMD/NVDA and can increase RMB FX volatility given China revenue sensitivity.

Risk assessment: Key tail risks are renewed China export controls or geopolitical restrictions (3–12 month horizon), a delay in AMD GPU deliveries to OpenAI (H2 2026 catalyst), or AI customers migrating to custom silicon (1–3 year). Immediate risk (days/weeks) is sentiment-driven multiple compression after a post‑earnings pop; short-term (quarters) risk is non-repeat of China revenue and semi‑custom decline; long-term (years) risk is competitive displacement by NVIDIA or in‑house chips. Hidden dependencies: 8 of top 10 AI firms using AMD GPUs creates concentration risk if a single large buyer shifts strategy.

Trade implications: Favor a core long in AMD sized 2–3% of portfolio, scaled over 2–4 weeks on pullbacks, with a 18–24 month hold to capture OpenAI H2 2026 deliveries and CPU share gains. Hedge 30% of that exposure with a 3‑month 15–20% OTM put spread to protect versus a non-repeat of China revenue. Add selective NVDA exposure via a 6–9 month call spread to play H2 AI ramp, and trim 2–4% cyclical gaming/console supply chain exposure now.

Contrarian angle: The market underweights quality of margin expansion (54% gross margin, +300bps) partly driven by MI308 reversal — a one‑time that may overstate operating leverage; yet forward PEG of 0.2 suggests expectations are conservatively priced vs growth, not expensive. Reaction to the China caveat looks overdone near-term, offering a buy-the-dip setup if Q1 guide holds; downside is forced multiple re-rating if OpenAI shipments slip, so keep tail hedges active through H2 2026.

Key Decisions for Investors

• Establish a 2–3% portfolio long position in AMD (AMD) over the next 2–4 weeks, scale in 2–3 tranches on any further pullback up to 15% from current levels; target an 18–24 month holding period to capture H2 2026 OpenAI GPU deliveries and CPU share gains, set tactical stop-loss at -20% from entry.
• Hedge 30% of the AMD position with a 3‑month put spread: buy a 15% OTM put and sell a ~35% OTM put (sized to 30% of the long notional) to protect against a non-repeat of the $390M China tail; be willing to pay up to ~2.5% of position value as premium.
• Allocate 1–2% portfolio to NVDA directional exposure via a 6–9 month call spread (buy ~20% OTM call, sell ~40% OTM call) to express H2 2026 AI infrastructure upside while capping premium outlay; take profits at ~2.5x premium or roll into longer-dated calls if OpenAI cadence confirmed.
• Reduce 2–4% exposure to gaming/console supply-chain names and semi-custom reliant suppliers over the next month (reallocate into AMD/NVDA or AI cloud names) given AMD’s guide for meaningful semi-custom revenue decline in 2026.