- AWS plans to add two million more Nvidia GPUs between 2027 and 2028.
- The expansion follows Amazon’s earlier commitment to deploy more than one million chips from 2026.
- New CPUs, faster networking and memory technologies will support larger AI workloads.
- Amazon is betting that demand for AI computing, robotics and advanced cloud services will keep growing.
Amazon is preparing for another major expansion of its artificial intelligence infrastructure, with AWS planning to deploy an additional two million Nvidia GPUs across its global data center network between 2027 and 2028.
The move reflects Amazon’s growing confidence that demand for AI computing will continue to rise, even as companies around the world pour unprecedented amounts of money into data centers, chips and high performance networking.
The latest commitment builds on an earlier plan to add more than one million Nvidia chips beginning in 2026. Together, the deployments represent one of the largest infrastructure expansions yet announced for cloud based AI computing.
AWS and Nvidia have worked together for years, but the scale of this new agreement shows how quickly the competition for AI capacity is intensifying. Businesses, governments and frontier AI companies are demanding larger and more powerful computing systems, particularly for training advanced models and running increasingly complex inference workloads.
Amazon appears determined to ensure AWS has enough capacity to meet that demand.
AWS Expands Beyond GPUs With New AI Infrastructure
The agreement is about considerably more than simply adding millions of GPUs.
AWS and Nvidia are also planning to introduce Nvidia’s Vera based CPUs into Amazon’s cloud infrastructure. These processors are expected to support the next generation of AI applications, particularly workloads that require close coordination between CPUs, GPUs, networking and memory.
The companies are also working on improvements to networking technology. NVLink Fusion and high bandwidth memory upgrades are expected to help large AI clusters process information more efficiently.
That matters because modern AI systems are no longer limited by raw computing power alone. As models become larger, the speed at which thousands of processors can communicate with one another becomes increasingly important.
A massive cluster of GPUs can still underperform if networking creates bottlenecks between systems. By expanding its investment across the full technology stack, AWS is attempting to build infrastructure capable of supporting the next wave of agentic AI, robotics and physical AI applications.
Amazon is also bringing Nvidia technology into other parts of its business.
Its analytics platforms are expected to benefit from Nvidia’s cuDF software, with some workloads approaching 3.7 times faster processing than conventional configurations. Amazon says these improvements could also provide around 30% better price performance compared with systems relying entirely on standard processors.
Search, Robotics and Government Computing Also Benefit
The GPU expansion will reach beyond traditional AI model development.
Amazon’s search technologies are receiving improvements through GPU acceleration, particularly for vector search workloads. Index construction is expected to become roughly nine times faster when GPUs are used.
Vector search has become increasingly important as companies build AI applications that need to understand relationships between large amounts of information rather than simply matching keywords.
Amazon’s robotics division is also working with Nvidia on simulation technologies designed to speed up the training of warehouse robots. Physical AI is becoming a major focus for the technology industry, with companies using powerful computing systems to train machines that can understand and interact with real world environments.
Government computing is another part of the plan.
Around 100,000 chips are expected to be allocated for sensitive government and defense workloads across the United States. That highlights the growing strategic importance of AI infrastructure, particularly as governments seek secure access to advanced computing resources.
AWS CEO Matt Garman said customers increasingly want flexibility when choosing AI technologies while also expecting those tools to work smoothly across cloud infrastructure.
Nvidia CEO Jensen Huang took an even broader view, describing the AWS and Nvidia relationship as one of the major growth engines of the AI era.
A Huge Bet on Future AI Demand
The biggest question surrounding Amazon’s investment is whether demand will continue growing quickly enough to justify such enormous infrastructure spending.
Companies across the technology industry are racing to secure GPUs and build data centers. Nvidia remains the dominant supplier of high performance AI chips, while cloud providers including Amazon are competing to provide customers with access to increasingly powerful systems.
Amazon’s decision to commit to another two million GPUs suggests it believes the current AI boom has much further to run.
The company is effectively betting that enterprises will continue adopting generative AI, that frontier labs will require larger computing clusters, and that entirely new categories of software and machines will emerge from advances in agentic and physical AI.
Recent hardware improvements have already produced significant gains. Newer systems are delivering substantially faster inference performance and stronger graphics capabilities than previous generations.
But expanding AI infrastructure at this scale also creates challenges.
Data centers require enormous amounts of electricity, advanced cooling systems and increasingly complex networking. Building enough facilities to support millions of additional high performance processors will require years of investment and careful planning.
For Amazon and Nvidia, however, the risk of not having enough computing capacity may be greater than the risk of building too much.
If AI demand continues to accelerate, companies with the largest and most capable infrastructure could hold a significant advantage.
Amazon’s latest agreement with Nvidia makes one thing clear: AWS is preparing for an AI future that will require computing power on a scale the industry has rarely seen before.
Follow TechBSB For More Updates
