Amazon's cloud division hit $37.6 billion in quarterly revenue while pouring nearly $200 billion into AI infrastructure, and the ripple effects are reaching decentralized compute markets.
Amazon Web Services just posted its best growth quarter in nearly four years, and the reason is exactly what you’d expect: everyone suddenly needs a lot more compute power to run AI workloads. The cloud giant reported Q1 FY2026 revenue of $37.6 billion, a 28% jump year-over-year, with record operating margins of 13.1%.
That’s the fastest AWS has grown in 15 quarters.
The numbers behind the AI gold rush #
AWS now carries an annualized revenue run rate of $150 billion. AWS’s revenue run rate for AI-related services has surpassed $15 billion.
Amazon has outlined capital expenditure plans approaching $200 billion for 2026, with the lion’s share directed at AI infrastructure. That includes new data centers, proprietary chips like Trainium and Graviton, and Project Rainier, a large-scale AI compute cluster designed to handle the most demanding workloads.
CEO Andy Jassy has called AI a “pivotal growth driver” for AWS.
Trainium and Graviton chipsets saw triple-digit growth in the quarter, suggesting that AWS customers are increasingly willing to use Amazon’s own hardware rather than defaulting to Nvidia.
Partnerships and the competitive landscape #
AWS has locked in key partnerships with AI firms, most notably Anthropic, and has a $6 billion multi-year commitment with Snowflake. Investments touching OpenAI and Meta further cement AWS’s position at the center of the AI infrastructure web.
The hyperscaler competition remains fierce. Microsoft Azure and Google Cloud are both pouring billions into similar AI buildouts.
What this means for crypto and decentralized compute #
AWS has started exploring Web3 services through platforms like its Bedrock offering, targeting use cases in decentralized finance, tokenization, and AI agents.
Tokens tied to decentralized GPU and compute services, like Render Network’s RENDER and Akash Network’s AKT, aggregate underutilized hardware from distributed sources, offering an alternative to renting capacity from AWS or Azure.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our