Uber, the on-demand mobility and delivery platform, has expanded its use of Oracle Cloud Infrastructure (OCI) to support its growing operations. According to the company, OCI now helps power more than one million Uber trips every hour and facilitates 14 million predictions per second through various AI models integrated into each ride.
In 2023, Uber selected Oracle to modernize its application tier and AI infrastructure. Since then, the company has migrated numerous microservices, data storage platforms, and AI models to OCI as part of its transition to cloud-based operations. The migration aims to enhance Uber’s ability to deliver new products, accelerate innovation, and manage infrastructure costs effectively.
Kamran Zargahi, Uber’s senior director of Tech Strategy and Cloud Engineering, emphasized that the collaboration with Oracle enables Uber to innovate more quickly while maintaining flexibility across multicloud environments. This approach is critical as Uber continues to scale its services and enter new markets.
Uber’s integration with OCI includes using OCI Compute with AMD to handle its trip-serving requests and OCI Compute with Ampere Arm for its stateless workloads. The company’s AI models, which power critical inferencing for its services, run on OCI’s AI infrastructure to optimize performance in terms of cost, throughput, and latency. Additionally, Uber has migrated part of its Hadoop Distributed File System (HDFS) and re-platformed its storage layer to OCI Object Storage, allowing for scalable, high-durability storage solutions.
