OpenAI is moving rapidly to diversify its computational resources following a recent corporate restructuring that granted the company greater flexibility in choosing cloud partners. The key to this new independence was waiving Microsoft’s first right of refusal to be OpenAI’s sole compute provider, a freedom the AI giant is now wasting no time in exercising.
On Monday, the company announced a monumental, multi-year cloud partnership with Amazon Web Services (AWS). This groundbreaking deal is valued at an astonishing $38 billion, significantly broadening OpenAI’s access to the immense computing power required to train and run its next-generation AI models.
Effective immediately, AWS will begin providing OpenAI with access to “thousands” of NVIDIA GB200 and GB300 GPUs. These powerful graphics processing units are critical for both the inference (running and operating models like ChatGPT) and the training of new, complex models.
Amazon’s plan is to deploy all the capacity that OpenAI has agreed to purchase by the end of 2026. The agreement also includes an option for OpenAI to secure the right to purchase additional cloud capacity from AWS in 2027 and beyond, cementing this as a long-term strategic alliance.
Amazon underscored the potential impact of the partnership, stating that the new resources “will help millions of users continue to get value from ChatGPT.” This highlights the scale of the infrastructure required not only for development but also for serving the massive global user base of OpenAI’s flagship products.
The sheer size of this AWS commitment, however, immediately raises questions about OpenAI’s financial outlay across its cloud providers. The company’s annualized revenue was recently reported to be approximately $12 billion.
This new Amazon deal is layered on top of the existing, massive agreement with its founding partner, Microsoft. As part of its restructuring, OpenAI previously agreed to spend a staggering $250 billion on Azure services from Microsoft. That agreement, along with a complex revenue-sharing deal tied to the potential development of Artificial General Intelligence (AGI), remains firmly in place.
The dual nature of these enormous cloud commitments—$250 billion with Microsoft and $38 billion with Amazon—signals OpenAI’s intent to operate across multiple hyperscalers, ensuring redundancy, negotiating leverage, and access to the best available hardware for its continued dominance in the rapidly evolving AI landscape.