Uber Expands AI and Cloud Infrastructure on AWS to Power Real-Time Global Operations

Uber  is expanding its cloud and artificial intelligence infrastructure on Amazon Web Services (AWS)  to enhance real-time ride matching, delivery efficiency, and AI-driven personalization across its global platform.

The company is increasing its use of AWS Graviton instances to power more of its Trip Serving Zones, the real-time systems responsible for handling every ride and delivery request. These systems process massive volumes of location data and make split-second decisions such as driver matching, route optimization, and arrival time predictions.

By leveraging Graviton-based compute, Uber aims to improve performance while reducing energy consumption and costs, particularly during peak demand periods such as rush hours and major global events. The upgrade is expected to enhance scalability, reduce latency, and improve system reliability across its mobility and delivery services.

In addition, Uber has begun pilot training some of its artificial intelligence models on AWS Trainium chips. These models help process billions of trip data points to optimize driver and courier assignments, improve delivery accuracy, and generate more precise estimated arrival times for users.

The company said the move will enable faster learning and smarter predictions across its platform, resulting in more personalized user experiences and improved operational efficiency.

According to Uber engineering leadership, the transition to AWS infrastructure provides the flexibility needed to handle global demand spikes while ensuring faster matching between riders, drivers, and delivery partners.

AWS leadership also noted that supporting Uber’s large-scale, real-time systems demonstrates the capability of its cloud and AI infrastructure to power next-generation mobility and on-demand services at global scale.