In a move that could redefine the global race for AI dominance, Amazon Web Services (AWS) and OpenAI have announced a sweeping, multi-year partnership worth $38 billion, positioning AWS as a key provider of the computing backbone behind OpenAI’s most advanced artificial intelligence systems.
Under the deal—one of the largest infrastructure agreements ever in the AI industry—OpenAI will immediately begin running and scaling its core AI workloads on AWS’s global network of data centers. The company will gain access to hundreds of thousands of NVIDIA GPUs, with plans to expand to tens of millions of CPUs over the next seven years to support increasingly complex “agentic” AI workloads.
The move comes as the explosive growth of generative AI drives an insatiable demand for computing power, forcing frontier model developers to seek ever-greater scale and reliability. “Scaling frontier AI requires massive, reliable compute,” said Sam Altman, OpenAI’s co-founder and chief executive. “Our partnership with AWS strengthens the broad compute ecosystem that will power this next era and bring advanced AI to everyone.”
AWS—long regarded as a pioneer in running secure and large-scale infrastructure—will deploy a customized compute architecture optimized for AI processing, featuring NVIDIA’s latest GB200 and GB300 GPUs clustered via Amazon EC2 UltraServers. This design allows OpenAI to run both inference workloads for ChatGPT and training for next-generation models with minimal latency and high efficiency.
For AWS, the partnership marks a high-profile endorsement at a time when competition in cloud computing is intensifying. Matt Garman, CEO of AWS, said the deal underscores the company’s unmatched ability to serve global-scale AI providers. “As OpenAI continues to push the boundaries of what’s possible, AWS’s best-in-class infrastructure will serve as a backbone for their AI ambitions,” he said. “The breadth and immediate availability of optimized compute demonstrates why AWS is uniquely positioned to support OpenAI’s vast AI workloads.”
The partnership builds on an existing collaboration between the two companies. Earlier this year, OpenAI’s open-weight foundation models became available on Amazon Bedrock, AWS’s platform for deploying generative AI models. Since then, thousands of customers—including Comscore, Peloton, Thomson Reuters, Triomics, and Verana Health—have begun using OpenAI’s models for applications ranging from scientific research and mathematical problem-solving to agentic workflow automation.
Industry analysts say the deal also signals a subtle shift in AI alliances. Microsoft remains OpenAI’s largest investor and cloud partner, but AWS’s new role as a major infrastructure provider adds complexity to the landscape. As OpenAI seeks to diversify its compute base, AWS gains a powerful validation of its technological leadership in the AI infrastructure space—potentially challenging Microsoft Azure’s dominance in the field.
Both companies are betting that scale will be the defining factor in the next chapter of artificial intelligence. By pairing OpenAI’s cutting-edge research with AWS’s mature global infrastructure, they aim to push the boundaries of what AI systems can learn, generate, and automate.
If successful, this partnership could not only accelerate the evolution of ChatGPT but also influence how governments, corporations, and individuals interact with machine intelligence in the years to come.