Amazon expands its reliance on Nvidia, committing to deploy an additional 2 million graphics processing units across AWS by 2028, signalling a significant shift in its AI infrastructure strategy amid rising industry demand.
Amazon is deepening its reliance on Nvidia as it expands the computing base behind its cloud and artificial intelligence services, committing to deploy another 2 million of the chipmaker’s graphics processing units across Amazon Web Services in 2027 and 2028. Bloomberg reported that the new order comes on top of the 1 million Nvidia chips Amazon said in March it would begin installing this year, underscoring how quickly the scale of its AI build-out is widening.
According to AWS and Nvidia, the latest expansion is intended to support a broad range of workloads, including agentic AI, enterprise automation, scientific research and physical AI. The companies also said the collaboration will bring Nvidia Vera CPUs to AWS as an additional compute option, while work on networking improvements is aimed at improving performance for large-scale AI training. AWS said the chips will be deployed across its global infrastructure.
The new commitment highlights the continued strength of Nvidia’s position in AI infrastructure, even as Amazon develops its own chips for data centres. TechCrunch reported that the planned deployment covers Nvidia’s Blackwell Ultra, Rubin and Rubin Ultra generations, reflecting the company’s push to secure capacity across multiple future platforms. The additional order also follows a period of unusually strong demand for AI hardware across the industry.
Taken together, the publicly disclosed commitments suggest Amazon is preparing for a far larger AI footprint than it outlined earlier in the year. TechTimes reported that the latest order lifts AWS’s disclosed Nvidia commitment to more than 3 million units over a roughly three-year period, though the company is still balancing that dependence with its own custom silicon strategy.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
