Amazon's AI Chips Win Uber Away From Oracle and Google in Cloud Infrastructure Fight

7. Amazon's AI Chips Win Uber Away From Oracle and Google in Cloud Infrastructure Fight

Uber is expanding its AWS contract to run ride-sharing features on Amazon's proprietary AI chips, a move that directly undercuts competing bids from Oracle and Google. The deal, reported by TechCrunch, signals that Amazon's silicon offering, built around its Trainium and Inferentia chip lines, is now competitive enough on price and performance to pull a major real-time compute workload away from rivals who have been aggressively courting enterprise AI infrastructure deals.

This matters because Uber's workload is not a vanity benchmark. Ride-sharing dispatch, surge pricing, and matching algorithms run at low latency against live data at massive scale, making it a credible proof point for Amazon's chips in production inference environments. Oracle has been packaging its GPU cloud as a cheaper NVIDIA alternative, and Google has been pushing its own TPUs through enterprise GCP deals. Losing Uber to AWS weakens both narratives. For Amazon, the win is doubly valuable: it deepens Uber's platform lock-in while generating a referenceable customer that other logistics and mobility companies will watch closely before their own contract renewals.

The broader pattern here is that the cloud hyperscalers are winning the AI infrastructure war not just on chips but on integration depth. AWS, Google, and Microsoft are each trying to make their proprietary silicon inseparable from their data, networking, and software tooling. Every enterprise that migrates a production AI workload onto a hyperscaler's custom chip becomes structurally harder to move. Uber's expansion is less about chips in isolation and more about which platform gets to own the full inference stack as AI becomes load-bearing infrastructure.

Source: https://techcrunch.com/2026/04/07/uber-is-the-latest-to-be-won-over-by-amazons-ai-chips/