Uber is expanding its use of Amazon Web Services’ custom chips to power real-time ride matching and train artificial intelligence models, the companies announced.
The ride-sharing giant is deploying AWS Graviton4 processors to support its Trip Serving Zones, the infrastructure behind every trip and delivery. This helps match riders with drivers in milliseconds, reduces energy consumption, and handles demand spikes during rush hours or major events, according to an official announcement by Amazon.
Uber also began piloting Trainium3 chips to train AI models that analyse billions of rides to improve arrival time estimates, driver assignments, and personalised recommendations, as per the official announcement.
The move expands an existing cloud partnership. Amazon is investing heavily in its custom chip business to attract enterprise customers as demand for AI training and inference grows.