OpenAI has signed a landmark $38 billion deal with Amazon Web Services (AWS) to access massive cloud computing capacity, marking a major shift from its former reliance on Microsoft’s Azure platform.
The agreement, announced Monday, makes Amazon the newest and most powerful partner in OpenAI’s expanding cloud ecosystem.
Under the multi-year deal, OpenAI will immediately begin using AWS’s infrastructure, powered by hundreds of thousands of Nvidia graphics processing units (GPUs) across the U.S., with plans to expand in the coming years.
Amazon’s stock rose roughly 5% following the announcement, underscoring investor enthusiasm about the tech giant’s growing role in artificial intelligence infrastructure.
AWS to Power OpenAI’s Next Frontier Models
The first phase of the partnership will leverage existing AWS data centers, while Amazon will build out new infrastructure tailored to OpenAI’s needs. “It’s completely separate capacity that we’re putting down,” said Dave Brown, vice president of compute and machine learning services at AWS. “Some of that capacity is already available, and OpenAI is making use of that.”
The deal will rely primarily on Nvidia’s latest Blackwell GPUs, with potential for other chips in the future. Although Amazon’s custom Trainium chips currently power Anthropic’s AI models, Brown declined to confirm whether OpenAI might adopt Trainium later, saying, “I can’t provide any details on anything we’ve done with OpenAI on Trainium at this point.”
Breaking Microsoft’s Exclusive Grip
For years, Microsoft held exclusive rights as OpenAI’s cloud provider, following its $13 billion investment since 2019. That exclusivity ended in January 2025, when Microsoft’s preferential status expired, allowing OpenAI to collaborate with other hyperscalers.
OpenAI has since struck agreements with Oracle, Google, and now AWS, which remains the global leader in cloud infrastructure. “Scaling frontier AI requires massive, reliable compute,” OpenAI CEO Sam Altman said. “Our partnership with AWS strengthens the broad compute ecosystem that will power this next era and bring advanced AI to everyone.”
Amazon Strengthens Its AI Cloud Dominance
For Amazon, the partnership underscores its strength in the competitive cloud market and highlights its expanding influence in AI. AWS CEO Matt Garman said, “The breadth and immediate availability of optimized compute demonstrates why AWS is uniquely positioned to support OpenAI’s vast AI workloads.”
Amazon’s cloud unit recently reported more than 20% year-over-year growth, outpacing analyst expectations, though still trailing Microsoft and Google’s 40% and 34% growth respectively. The deal with OpenAI is expected to further cement AWS’s lead in high-performance cloud computing.
OpenAI’s IPO Path and AI Infrastructure Expansion
OpenAI continues to spend heavily on cloud infrastructure, reaffirming last week that it will purchase an additional $250 billion in Azure services. Yet, its growing network of partners signals a move toward independence and operational maturity, crucial steps as it prepares for a possible initial public offering.
Altman recently said in a livestream that an IPO is “the most likely path” for OpenAI, while CFO Sarah Friar described the company’s restructuring as preparation for public markets. “As part of this deal, OpenAI is a customer of AWS,” Brown said. “They’ve committed to buying compute capacity from us, and we’re charging OpenAI for that capacity. It’s very, very straightforward.”
AI Ecosystem Broadens as Compute Wars Intensify
The infrastructure secured under the deal will support both inference, such as real-time ChatGPT responses, and training of future AI models. OpenAI’s foundation models, including open-weight options, are already available on AWS’s Bedrock platform, which also powers applications at companies like Peloton, Thomson Reuters, and Comscore.
As the $500 billion startup expands its cloud footprint across hyperscalers, the new pact with Amazon marks a pivotal moment in the global race for AI compute dominance. It also signals that OpenAI’s ambitions, once tethered to a single partner—are now distributed across an increasingly competitive AI ecosystem.
