OpenAI announced a $38 billion agreement with Amazon on Monday to purchase cloud computing services over seven years, enabling the company to scale its AI infrastructure following a recent restructuring that removed the need for Microsoft’s approval on such deals.
The deal allows OpenAI to begin utilizing Amazon Web Services compute resources immediately. All purchased capacity under the agreement is scheduled for deployment before the end of 2026. OpenAI has the option to extend its usage of AWS into 2027 and subsequent years, supporting ongoing expansion of its computational capabilities.
This partnership with Amazon forms a key component of OpenAI’s overarching strategy to significantly increase its computing power. The company anticipates spending more than $1 trillion on such infrastructure over the next decade to meet the demands of developing advanced AI systems, including agentic workloads that power tools like ChatGPT.
In parallel efforts, OpenAI has initiated new data center construction projects with multiple partners, including Oracle, SoftBank, and the United Arab Emirates, along with additional collaborators. These buildouts aim to bolster the physical infrastructure necessary for large-scale AI operations.
OpenAI has also finalized agreements with leading chip manufacturers Nvidia, AMD, and Broadcom to secure essential hardware components. These deals ensure access to high-performance processors critical for training and running complex AI models.
