Uber is expanding its infrastructure and AI capabilities using technology from Amazon Web Services (AWS), it announced on Tuesday.

The move will see Uber run more of its Trip Serving Zones, systems that match riders with drivers and manage delivery logistics, on AWS infrastructure powered by AWS’s Graviton4 processors. The company said the shift will reduce latency, improve scalability during demand spikes and lower energy consumption.

Trip Serving Zones power Uber’s real-time operations by processing location data in milliseconds to determine driver availability, estimated arrival times and routing decisions.

By expanding its use of AWS processors, Uber said it aims to accelerate these calculations during demand spikes while optimising cost and performance by reducing its energy consumption.

Alongside infrastructure changes, Uber is piloting AI model training on AWS Trainium3, a custom chip designed for large-scale machine learning workloads. The models analyse data from billions of rides and deliveries to improve rider-driver matching, demand forecasting and personalised recommendations.

Uber said training AI systems at this scale requires significant computing power and that the use of Trainium3 will enable faster training cycles and more efficient model development. As models improve, the company expects more accurate arrival time predictions, smarter routing and a better user experience across its platform.

"Uber operates at a scale where milliseconds matter," said Kamran Zargahi, vice president of engineering at Uber. "Moving more Trip Serving workloads to AWS gives us the flexibility to match riders and drivers faster and handle delivery demand spikes without disruption.”

“By starting to pilot some of our AI models on Trainium,” he added, “we are building a technology foundation that will make every Uber experience smarter so we can keep our focus where it belongs: on the people who use Uber every day.”

---