AMD (AMD) Q3 2026 Earnings Call Transcript

AMD reported record Q3 2026 revenue of $13.4 billion, up 45% year over year, and non-GAAP EPS of $2.09, up 74%; data center revenue rose 97% to $8.5 billion. Management forecast Q4 revenue of approximately $17 billion at the midpoint and FY2027 revenue growth above 80%, with server revenue expected to double and data center AI revenue to grow even faster, while noting demand exceeds available supply. Offsetting risks include a projected double-digit percentage decline in 2027 PC market unit shipments, higher component costs, and variable gross margins as data center mix shifts.

Analysis

AMD’s upside is increasingly a capacity-allocation and execution story, not simply an AI-demand story. If GPU supply remains constrained, its server CPU ramp can still monetize adjacent agent and accelerator-hosting demand—but the company may not capture all of the upside implied by customer plans. For the GPU ramp, rack-level performance and software reliability must translate into repeat deployments; customer announcements and management’s TCO claims are not yet independent proof of durable share. Watch shipment conversion and customer concentration.

The key second-order tension is mix: faster accelerator growth can lift profit dollars while diluting consolidated gross margin, particularly if HBM and packaging costs rise faster than customer pricing adjusts. Inventory build and lower free-cash-flow conversion make shipment timing and working-capital normalization important, even with reported operating leverage. Separately, expected warrant remeasurement can make GAAP earnings noisy; it is noncash, but eventual dilution conditions and share count still warrant monitoring.

Near term, the Q4 ramp creates upside if shipments and margins execute; disappointment in either can unwind elevated growth expectations quickly. Over 1–3 months, verify data-center revenue conversion, gross-margin trajectory, inventory turns and cash conversion. Over 6–18 months, software adoption, supply availability and repeat enterprise deployments determine whether AMD earns a sustained share shift. The contrarian risk is that strong demand forecasts are being treated as equivalent to profitable, deliverable demand; the opportunity is that CPU growth may be underappreciated alongside accelerators.

Key Decisions for Investors

• Favor AMD on a post-event consolidation rather than chasing an opening spike; consider a staged long, with risk defined below the post-earnings low. The thesis is shipment execution and server-CPU share gains; reduce or exit if data-center growth decelerates materially or gross-margin deterioration persists beyond mix-related quarterly variation.
• For a relative-value expression, pair a modest long AMD with a short HP Inc. to offset some PC-market exposure while emphasizing AMD’s data-center mix. Keep the hedge smaller than the AMD leg: this is not a clean hedge, and it can lose if AMD’s valuation compresses or HP Inc. outperforms for company-specific reasons.
• Over the next 1–3 months, track data-center shipments, segment operating margin, inventory growth versus revenue, and free-cash-flow conversion. Treat continued inventory accumulation without corresponding shipment growth—or evidence of weaker customer deployments—as an alert to cut exposure.
• Do not underwrite the thesis on headline non-GAAP EPS alone. Reconcile GAAP results for warrant remeasurement, then monitor disclosures on vesting/exercise conditions and diluted share count; a worse-than-expected dilution path would weaken per-share upside even if operating demand remains strong.