Amazon has said it will integrate Nvidia chips in its new Amazon Web Services (AWS) AI factories.

The AI factories are a new offering designed to provide enterprises and governments with dedicated AWS AI infrastructure deployed in their own data centres.

They aim to help businesses and public sector organisations meet their data sovereignty and regulatory requirements, with accelerated implementation times.

Amazon said that the factories will enable customers to leverage their existing data centre space, network connectivity, and power while AWS handles the complexity of deployment and management of the integrated infrastructure.

Through the partnership with Nvidia, AWS customers will get access to the Nvidia accelerated computing platform, full-stack Nvidia AI software, and thousands of GPU-accelerated applications to build AI applications.

Amazon explained the move comes at a time when governments and large organisations are looking to expand AI projects and use AI factories to meet their specific sovereignty and compliance needs.

However, these organisations are unable to tackle the challenges of creating a high-performance AI factory that requires a vast set of highly complex management, database, storage and security services on their own, Amazon continued.

“Building their own AI capabilities requires massive capital investments in GPUs, data centres and power, plus navigating complex procurement cycles, selecting the right AI model for their use case, and licensing models from different AI providers,” the company added. “This creates a multi-year timelines and operational complexity that diverts focus from their core business goals.”

AWS AI Factories will combine a wide range of AI accelerators, including “cutting-edge” Nvidia AI computing and Trainium chips, AWS high-speed, low-latency networking, storage and databases, security, and infrastructure.

The stack will also include Amazon’s AI services, such as Amazon Bedrock and SageMaker, to further support customers in developing and deploying AI applications at scale.

Ian Buck, vice president and general manager of Hyperscale and HPC at Nvidia, emphasised how large-scale AI requires a full stack approach which includes advanced GPUs and networking, as well as software and services that optimise every layer of data centres.

He added: “By combining Nvidia’s latest Grace Blackwell and Vera Rubin architectures with AWS’ secure, high-performance infrastructure and AI software stack, AWS AI Factories allow organisations to stand up powerful AI capabilities in a fraction of the time and focus entirely on innovation instead of integration.”

---