Competition Is Heating Up, But Bank of America Still Thinks AMD Stock Is a Buy Here

AMD reported fiscal Q3 2025 revenue of $9.2 billion (+36% YoY) and adjusted EPS of $1.20 (+30% YoY), beating estimates, with data-center revenue of $4.34 billion (+22%) and Client & Gaming at $4.0 billion (+73%). Management guided fiscal Q4 revenue around $9.6 billion (+/- $300 million; midpoint ~25% YoY) and non-GAAP gross margin near 54.5%, while securing a major multi-year OpenAI agreement (6 GW of Instinct GPUs, 1 GW initial rollout) that strengthens its AI positioning versus Nvidia. The story is tempered by emerging competitive risk after reports Meta may invest in Google TPUs — a development that briefly pressured AMD shares — but analysts including Bank of America remain constructive.

Analysis

Market Structure: AMD is a near-term beneficiary of multi‑vendor demand as OpenAI’s 1GW initial rollout (H2 next year) and a potential multi‑GW multi‑year commitment expand TAM; Bank of America’s $1.2T AI data‑center TAM by 2030 implies room for multiple winners. Hyperscalers building custom TPUs (Google→Meta) subtract high‑margin internal GPU spend but likely won’t displace public cloud GPU demand (AWS/MSFT/AMZN) in the next 12–36 months, preserving pricing power for AMD/ Nvidia where HBM and TSMC capacity remain constrained.

Risk Assessment: Tail risks include a confirmed Meta→Google TPU shift at scale (2027+, material revenue loss >5% CAGR for GPUs), adverse export controls on AI accelerators, and foundry/HBM supply shocks that could compress AMD’s guided ~54.5% gross margin. Immediate (days) downside is headline‑driven; short term (3–6 months) depends on Q4 execution and OpenAI rollout signals; long term (2–5 years) hinges on software ecosystem (ROCm vs CUDA) and TSMC node availability.

Trade Implications: Tactical overweight AMD (stock or defined‑risk calls) to capture OpenAI optionality while sizing for headline volatility; pair trades long AMD vs short NVDA to capture relative valuation re‑rating if AMD continues share gains, keep net exposure market‑neutral. Buy Jan 2026 AMD call spreads (limited risk ≤1% NAV) and add selective exposure to semiconductor equipment (ASML) and HBM/memory beneficiaries (MU) to play structural capex growth.

Contrarian Angles: Consensus underweights the supply‑side constraints and software lock‑in; a confirmed Meta TPU deal is a slow, high‑capex transition (2027+), so near‑term selloffs are likely overdone if AMD hits guidance. Historical parallels (Intel/AMD server share shifts) show software/ecosystem inertia; hyperscaler custom chips can paradoxically expand total industry capex, benefiting foundries and GPU competitors over 2–4 years.

Key Decisions for Investors

• Establish a 2–3% long position in AMD (AMD) equity over the next 30 trading days, scaling in on any pullback ≥10%; target horizon 9–18 months to capture OpenAI deployment and expected share gains, stop‑loss at 12% from entry or if Q4 guide misses midpoint by >3%.
• Purchase a Jan 2026 AMD bull call spread (risk‑defined) sized at 0.5–1.0% NAV (e.g., buy ATM-ish call, sell higher strike ~+20–30%) to leverage upside to 2030 TAM while capping premium; unwind if AMD outperforms NVDA by >25% or OpenAI deployment stalls for >6 months.
• Implement a relative value pair: long AMD 2.0% vs short NVDA 1.5% (dollar‑neutral weighting) to play potential mean reversion; reduce short if NVDA reports a data‑center beat or if AMD/NVDA relative widens >15% in 60 days.
• Rotate 1–2% into semiconductor equipment (ASML) and HBM/memory beneficiaries (Micron MU) over 3–12 months to capture upstream capex growth; trim discretionary ad/consumer tech exposure (reduce META by 1–2%) if Meta confirms large TPU adoption within 90 days.