# Uber has expanded its AWS contract to run real-time ride-matching infrastructure on Amazon’s Graviton4 processor and is piloting AI model training on.

*genai · news · 2026-04-08 · TNW*

## Key points

- Uber is moving its real-time Trip Serving Zones infrastructure to AWS Graviton4 processors for improved latency.
- Uber is piloting AI model training on Amazon's Trainium3, leveraging its massive proprietary trip dataset.
- Trainium3 offers 30–50% lower costs than Nvidia H100/H200 for large-scale AI training workloads.
- Uber now uses all three major hyperscalers—AWS, Google Cloud, and Oracle Cloud—for different workloads.
- Uber's adoption of Trainium3 expands Amazon's customer base beyond AI labs to high-frequency operational workloads.

**Companies:** Uber, Amazon, Apple, OpenAI, Anthropic
**Countries:** United States

[Read the full story on TNW](https://thenextweb.com/news/uber-amazon-trainium-aws-chips)

---

Canonical: https://newsio.io/n/f7960508-8bd3-4628-9905-1b3f9c478c47/uber-has-expanded-its-aws-contract-to-run-real-time-ride-matching-infrastructure
Summarized by Newsio from TNW. https://newsio.io/how-it-works