Uber Joins Amazon’s Trainium Roster with AWS Expansion Deal
April 8, 2026 – 9:24 am
In short:
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 Trainium3. This joins a growing list of major tech companies leveraging Amazon’s custom silicon strategy, highlighting its effectiveness.
Uber’s system runs on milliseconds, with Trip Serving Zones determining driver matches in real time before the user even finishes loading. Scaling this operation to 40 million trips daily across 72 countries requires substantial compute power and zero latency tolerance.
The latest deal sees Uber moving this critical workload to AWS, utilizing Graviton4 for Trip Serving Zones, and beginning a pilot to train AI models on Trainium3 using historical trip data.
What Uber is Moving and Why:
The announcement encompasses two key areas:
- Trip Serving Zones (real-time infrastructure): This component, not involving generative AI, requires high responsiveness under heavy load. Graviton4’s design for high-throughput, low-latency compute makes it ideal for this purpose.
- AI Model Training: Uber is leveraging its vast dataset of 13.567 billion trips and 200 million monthly active users to train AI models on Trainium3. The cost-effectiveness of Trainium3 makes the pilot a financially sound decision, even before considering potential performance gains.
Kamran Zargahi, Uber’s vice president of engineering, emphasized:
"Uber operates at a scale where milliseconds matter… Moving more Trip Serving workloads to AWS gives us the flexibility to match riders and drivers faster and handle delivery demand spikes without disruption."
And Rich Geraffo, vice-president and managing director for North America at AWS, highlighted Uber’s real-time demands:
"Uber is one of the most demanding real-time applications in the world…"