Anthropic growth set to boost Amazon's AWS revenue acceleration, says Bank of America

Bank of America reiterates a Buy on Amazon with a $275 target, citing Anthropic's rapid ARR ramp to roughly $19 billion from $9 billion in December as a potential catalyst for stronger-than-expected AWS revenue growth. Analysts estimate Anthropic’s surge implies more than $2.5 billion in quarterly revenue for Anthropic and say that if a significant share of its workloads run on AWS it could add up to $1 billion in quarter-over-quarter AWS revenue in Q1—above their ~$900 million baseline growth estimate. They also highlight Anthropic’s projected hyperscaler payments rising to as much as $6.4 billion in 2026 (from $1.9 billion in 2025) and note Amazon’s plan to double AWS power capacity by 2027 as supporting faster revenue traction and improved returns on related capex.

Analysis

Market structure: AWS and GPU suppliers (NVDA) are the clear near-term winners if Anthropic's ARR and cloud spend are real—Bank of America’s note implies up to a $1B quarter‑over‑quarter lift to AWS and an incremental $6.4B of hyperscaler revenue in 2026 vs $1.9B in 2025. Smaller clouds and on‑prem vendors (DELL, HPE) are structurally disadvantaged as AI workloads concentrate with hyperscalers; expect pricing power to tilt toward providers who can supply scale and specialized accelerators.

Risk assessment: Key tail risks include host‑concentration (Anthropic may split workloads across AWS/Azure/GCP), regulatory action on revenue‑sharing or export controls, or rapid migration to custom silicon that reduces hyperscaler spend; each could flip upside to downside within 1–12 months. Immediate window (days–weeks) is sensitive to hosting-disclosure headlines; short term (Q1–Q2) to reported AWS revenue beats >$900M; long term (through 2027) to Amazon’s capacity expansion execution and capex returns.

Trade implications: Direct play is AMZN exposure to capture AWS upside and NVDA exposure for accelerator demand; use defined‑risk call spreads around near‑term earnings and 6–18 month view on capacity doubling. Pair opportunities: long AMZN vs short on‑prem vendors (DELL) to express secular shift; rotate sector weight toward Cloud, Semis, Data‑center REITs only if AWS revenue beats thresholds (> $1B incremental from Anthropic).

Contrarian angles: Consensus assumes AWS captures most Anthropic workloads — that’s unproven and contract terms (revenue share, margin splits) could compress hyperscaler margins materially. Historical parallels (large single‑customer cloud relationships) show outsized volatility; if Anthropic diversifies hosts or builds private infra, current enthusiasm is overdone and AMZN upside will be limited despite ARR headlines.

Key Decisions for Investors

• Establish a 2.5% long position in AMZN using a defined‑risk Jun 2026 call spread (buy Jun 2026 240C / sell Jun 2026 320C) to capture potential AWS upside into Q2; if AWS reports an incremental >$1.0B QoQ revenue contribution tied to Anthropic, increase AMZN exposure to 5% within 2 trading days; cut to zero if AWS growth tied to AI is < $500M QoQ.
• Implement a relative‑value pair: long AMZN 2.0% / short DELL 1.25% (equal dollar) over a 6–12 month horizon to express cloud share gains vs on‑prem decline; unwind if DELL outperforms AMZN by >10% over any 60‑day window or if DELL guidance is upgraded materially.
• Buy a 1.0% portfolio‑weight NVDA 3–6 month call spread (defined strikes to limit premium) to capture GPU demand from Anthropic/OpenAI scale; trim if NVDA data‑center revenue guidance misses by >10% or if AI server orders decelerate sequentially.
• Trim enterprise hardware exposure (DELL/HPE/CSCO) by 1–2% and redeploy into Cloud/Semiconductor winners if within 30 days public disclosures show >50% of Anthropic workloads on AWS or if market consensus raises AWS revenue estimates by ≥$900M for Q1.