Datamation content and product recommendations are
editorially independent. We may make money when you click on links
to our partners.
Learn More
Amazon has struck a $38 billion partnership with OpenAI that immediately grants the ChatGPT maker access to hundreds of thousands of Nvidia GPUs.
This is an infrastructure deal that could reshape how AI reaches every corner of the economy.
With OpenAI projecting revenue growth from $13 billion to $125 billion by 2029, the strategy is plain: break free from single-cloud dependence and scale beyond what any one provider can deliver.
OpenAI began using AWS infrastructure under a seven-year agreement, according to Amazon’s announcement. OpenAI gets access to thousands of Nvidia GPUs, and can expand to tens of millions of CPUs as workloads spike.
The architecture stacks up with intent, clustering both GB200 and GB300 NVIDIA GPUs via Amazon EC2 UltraServers for low-latency performance across interconnected systems. All planned capacity will be deployed by the end of 2026, with room to grow through 2027 and beyond.
AWS brings scale and a track record for running AI systems. The company operates large clusters exceeding 500,000 chips, a level of experience that matters when you are pushing training and inference around the clock.
Pleasant partnership
This tie-up changes how AI reaches businesses and consumers. OpenAI is already one of the most popular model providers in Amazon Bedrock, serving customers including Peloton, Thomson Reuters, and Verana Health for everything from agentic workflows to scientific analysis.
The move also shows OpenAI’s diversification. While Microsoft invested $1 billion in 2019 and another $2 billion in 2021 to become OpenAI’s exclusive cloud provider, the AWS deal signals a shift toward multi-cloud infrastructure.
Future of AI
This partnership positions both companies for the next phase of AI, where infrastructure is as critical as the models. It gives OpenAI the headroom to run ChatGPT inference, then turn around and train next-generation models.
For businesses, it means easier access. OpenAI’s capabilities flow through AWS’s enterprise channels, so teams can plug models into applications via Amazon Bedrock and SageMaker. Think financial forecasting, customer service automation, the everyday tools people already use.
The ripple effects run wider than two companies. This partnership accelerates the shift to cloud-based AI infrastructure. And if AI agents generate around $29 billion in revenue for OpenAI by 2029, the pipes underneath become the real story, the backbone for a new wave of intelligent applications.