Amazon Web Services (AWS) and OpenAI have announced a multi-year strategic partnership valued at $38 billion to expand OpenAI’s computing capacity and support its growing AI workloads. The collaboration gives OpenAI immediate access to AWS’s large-scale cloud infrastructure, including hundreds of thousands of NVIDIA GPUs and the ability to scale to tens of millions of CPUs by 2027.
The partnership is designed to meet the rising demand for computing power needed to train and operate advanced generative AI systems. AWS will provide OpenAI with Amazon EC2 UltraServers, clustering NVIDIA GB200 and GB300 GPUs to enable low-latency and high-performance AI processing. This infrastructure will power applications such as ChatGPT and the training of next-generation models.
Both companies emphasized the importance of reliable and scalable compute for the next stage of AI development. Earlier in 2025, OpenAI models became available through Amazon Bedrock, allowing AWS customers to integrate OpenAI’s tools for coding, data analysis, and automation across industries.
Leave a comment