OpenAI has signed a $38 billion multi-year cloud deal with Amazon Web Services (AWS) — its first major contract outside Microsoft Azure. The move comes days after OpenAI renegotiated its partnership with Microsoft, removing the clause that forced it to use Azure exclusively.
The agreement gives OpenAI access to massive GPU and CPU capacity across AWS data centers, marking a decisive turn toward multi-cloud independence as competition for compute power heats up.
“Scaling frontier AI requires massive, reliable compute,” said CEO Sam Altman. “Our partnership with AWS strengthens the ecosystem that will power this next era.” AWS CEO Matt Garman called it proof that Amazon’s infrastructure “is uniquely positioned to support OpenAI’s vast AI workloads.”
From Azure Lock-In to Cloud Freedom
Under the deal, OpenAI will spend $38 billion on AWS compute through 2026, including EC2 UltraServers equipped with Nvidia GB200/GB300 GPUs. The capacity will train next-gen GPT models and serve ChatGPT at scale.
Freed from Azure exclusivity, OpenAI can now use AWS, Oracle, Google, and CoreWeave. It still plans to spend another $250 billion on Azure, signaling partnership rather than breakup — but the balance of power has clearly shifted.
Amazon’s Strategic Victory
For Amazon, hosting OpenAI alongside existing partner Anthropic is a major coup. AWS is building an $11 billion Indiana campus for Anthropic, and now adds OpenAI’s workloads — effectively making it the “AI supermarket” of cloud computing.
The deal intensifies the AI compute arms race, where hyperscalers rush to secure scarce Nvidia chips. OpenAI’s leadership has long warned that insufficient compute is its biggest threat; this contract locks in a vast, guaranteed supply.
The OpenAI-AWS alliance redefines the AI infrastructure map.
Freed from Azure exclusivity, OpenAI is betting that diversified, multi-cloud capacity is the only way to stay ahead in the race to build the next frontier of intelligence — and Amazon just scored one of the biggest wins in cloud history.
