Amazon has announced that Uber is expanding its agreement with Amazon Web Services (AWS) to run more of its platform on Amazon’s in-house AI and compute chips. The expansion includes increased use of Graviton, AWS’s Arm-based processors, and a new trial of Trainium, its AI training chip positioned as a competitor to Nvidia.
The move marks a shift in Uber’s cloud strategy. While the company had committed to migrating infrastructure to Google Cloud and Oracle in 2023, it is now increasing reliance on AWS, particularly for AI-driven workloads. Amazon is positioning its custom silicon as a key differentiator, attracting major customers seeking alternatives to traditional chip providers.
The deal underscores intensifying competition in AI infrastructure, with AWS leveraging vertically integrated hardware to win enterprise workloads. Uber joins companies such as Anthropic, OpenAI, and Apple in expanding usage of AWS chips as demand for AI compute continues to accelerate.
