Amazon increases CAPEX to $44.2B in Q1, accelerating AI buildout

Per Quartz, Amazon reported quarterly capital expenditures of $44.2 billion in Q1 2026, up 77% from $25 billion in Q1 2025 (Quartz). CryptoBriefing reports Amazon's trailing twelve-month capex reached $147.3 billion, a 67% increase from $88 billion in the prior twelve-month period (CryptoBriefing). Quartz and CryptoBriefing both report that trailing twelve-month free cash flow fell to $1.2 billion, down sharply from prior-period levels (Quartz; CryptoBriefing). Quartz also reports AWS revenue rose 28% year over year to $37.6 billion for the quarter (Quartz).

Reporting by Yahoo Finance and 247WallSt cites commitments backing the spend: an AWS backlog of about $364 billion, including over $225 billion in Trainium commitments and more than $100 billion tied to Anthropic, according to those outlets (Yahoo Finance; 247WallSt). Quartz reports that Amazon disclosed a chips business (Graviton/Trainium/Nitro) above a $20 billion annual run rate and that the company processed more tokens through Bedrock in Q1 2026 than in all prior years combined, with customer spend on the platform growing 170% quarter-over-quarter (Quartz). Quartz also states Q1 net income included pre-tax gains of $16.8 billion from investments in Anthropic (Quartz).

Large-scale AI infrastructure programs commonly require concentrated capital outlays for data centers, networking, and custom silicon. Companies reporting similar expansions often cite internal accelerator chips and platform services as levers to lower per-inference costs while increasing margin capture. For practitioners, that typically translates into greater public-cloud capacity for high-throughput model training and inference, more on-ramps for procurement of Trainium-style capacity, and faster iteration cycles for large-model deployments.

Editorial analysis: For the broader market, a quarter with $44.2 billion in CapEx and a reported $200 billion annual program frames Amazon as a major capacity builder in the AI compute layer. That scale of spending compresses near-term free cash flow, as reported, while potentially shifting the competitive landscape for cloud compute economics if customers commit to long-term capacity purchases. Observers covering cloud economics and model deployment will view the combination of high CapEx and growth in Bedrock and chip revenue as a signal that sizeable cloud providers are turning from incremental to structural infrastructure investments.

Editorial analysis: Observers should monitor these indicators over coming quarters:

• •AWS revenue growth relative to capex cadence and monetization of new capacity
• •the pace of Trainium and chips revenue growth versus reported run rates
• •changes in free cash flow as new capacity enters service and customer commitments convert to realized revenue. Reporting also flags customer commitments and backlog figures as key inputs that market participants are using to judge ROI on the large capex program (Yahoo Finance; 247WallSt)