US ride-hailing platform Uber has announced a partnership with Amazon Web Services (AWS) to deploy its in-house custom chips, aiming to improve the speed and efficiency of artificial intelligence (AI) model training and inference. The move is expected to strengthen Uber's real-time computing capabilities across ride-hailing and delivery services, while highlighting a broader push by cloud providers to capture AI infrastructure demand through proprietary silicon.
According to Reuters, Uber plans to adopt two core AWS chip architectures. The first is the high-performance, energy-efficient AWS Graviton processor, which will support day-to-day operations such as real-time ride matching, route optimization, and logistics systems. The second is AWS Trainium, designed specifically for large-scale machine learning workloads, enabling faster AI model training. The combination allows Uber to balance cost and performance while reducing training cycles.
At the application level, increased AI computing power is expected to translate directly into improved user experience. Uber said the new architecture will accelerate driver-passenger matching and enhance personalization through data-driven insights, including demand forecasting, dynamic pricing optimization, and delivery efficiency improvements, strengthening its position in the global ride-hailing and food delivery markets.
For AWS, the partnership carries strategic weight. The company has been expanding its in-house chip ecosystem to reduce reliance on traditional GPU vendors while offering more cost-competitive solutions to enterprise customers. With demand for generative AI and large language models (LLMs) surging, enterprises are seeking high-performance, scalable, and cost-efficient computing resources, making custom silicon a key differentiator for cloud service providers.
Analysts noted that Uber's adoption of AWS chips reflects more than a single-company upgrade, pointing instead to a broader industry shift from general-purpose computing toward vertically integrated, application-specific architectures. This transition suggests that cloud platforms are evolving beyond infrastructure providers into full-stack AI solution vendors that combine hardware and software capabilities.
For Uber, the move also supports risk diversification and cost control, helping reduce long-term computing expenses while improving system stability and scalability.
Article translated by Levi Li and edited by Jack Wu
