OpenAI Signs $38 Billion Deal With Amazon to Access Nvidia GPUs for AI Expansion as Microsoft Exclusive Rights End

OpenAI Taps Amazon for $38 Billion Cloud Deal to Power AI Ambitions

OpenAI is turning to Amazon Web Services (AWS) in a major move to fuel its rapidly expanding artificial intelligence projects, agreeing to a seven-year, US$38 billion (S$50 billion) deal to access hundreds of thousands of Nvidia graphics processors.

The agreement, announced on 3 November, follows a recent restructuring that gave the ChatGPT developer greater operational and financial independence and reduced its reliance on Microsoft.

How Will AWS Support OpenAI’s AI Expansion

The partnership will provide OpenAI with immediate access to AWS infrastructure, with full planned capacity expected by the end of 2026 and further room to grow in 2027 and beyond.

Amazon will deploy hundreds of thousands of Nvidia GB200 and GB300 AI accelerators in dedicated data clusters designed to optimise AI training and inference.

Dave Brown, AWS Vice President of Compute and Machine Learning Services, said,

“It’s entirely separate capacity that we’re putting down. Some of that capacity is already available, and OpenAI is making use of that.”

OpenAI CEO Sam Altman explained the reasoning behind the deal:

“Scaling frontier AI requires massive, reliable compute. Our partnership with AWS strengthens the broad compute ecosystem that will power this next era and bring advanced AI to everyone.”

Altman has previously outlined plans to invest US$1.4 trillion in developing 30 gigawatts of computing power—enough to supply roughly 25 million US homes—and has indicated ambitions to add one gigawatt of compute every week, each gigawatt carrying a capital cost of over US$40 billion.

How OpenAI Is Moving Beyond Microsoft

The AWS deal comes just days after OpenAI altered its relationship with Microsoft, which first invested US$13 billion in the company in 2019.

Under new commercial terms, Microsoft’s exclusive cloud rights ended, allowing OpenAI to expand partnerships with other hyperscalers.

OpenAI has also signed agreements with Google and Oracle, though AWS remains the largest cloud provider.

The shift signals OpenAI’s strategy to diversify infrastructure while maintaining a global AI expansion.

Analysts view the deal as a major endorsement for AWS.

Paolo Pescatore from PP Foresight said,

“This is a hugely significant deal (and is) clearly a strong endorsement of AWS compute capabilities to deliver the scale needed to support OpenAI.”

What This Means for AWS and the AI Industry

The announcement immediately lifted Amazon shares, closing 4 per cent higher on 3 November after a near 10 per cent jump on 31 October.

The deal bolsters AWS’s position in the AI market at a time when investors feared it had fallen behind Microsoft and Google.

Amazon’s infrastructure for OpenAI is designed to handle the demanding workloads of AI research and ChatGPT operations.

Clustering Nvidia GPUs on Amazon EC2 UltraServers enables low-latency performance, supporting training of next-generation models and delivering inference for users in real time.

How OpenAI Plans to Expand Global AI Infrastructure

OpenAI has been actively building its computing footprint, announcing agreements worth around US$1.4 trillion with Nvidia, Broadcom, Oracle, and Google.

Its data centre projects, collectively called Stargate, aim to deliver 30 gigawatts of computing power.

This includes five additional US data centres, bringing existing capacity to about 7 gigawatts, with plans to invest over US$400 billion in AI infrastructure over the next three years.

Oracle delivered the first Nvidia GB200 racks in June, and OpenAI has already begun early training workloads at its flagship Texas location in Abilene.

Expansion continues with new facilities in Shackelford County, Texas, and Stargate Argentina, signalling a significant scaling of AI infrastructure worldwide.

Is the AI Boom Fuelled by a Trillion-Dollar Appetite for Compute

The AWS deal highlights the unprecedented demand for computing power as AI systems grow more complex.

While OpenAI’s ambitious spending and infrastructure plans point to long-term AI dominance, some industry observers have raised concerns about the scale of investment, cautioning that the surge in valuations and spending could indicate a bubble in the AI sector.

By leveraging AWS’s global cloud infrastructure and Nvidia’s high-performance GPUs, OpenAI positions itself to accelerate AI research, train next-generation models, and expand access to advanced AI capabilities worldwide.