Cloud provider CoreWeave has inked a multi-year agreement with Wayve for building embodied intelligence in its AI Driver tech for autonomous vehicle use cases.
Founded in 2017 and backed by the likes of SoftBank Group, Nvidia, Uber and Eclipse Ventures, Wayve has a stated mission to “reimagine mobility and beyond” with embodied intelligence, “reshaping the future of autonomous mobility” with embodied artificial intelligence (AI), encompassing assisted to automated driving.
Its Embodied AI technology is at the heart of its foundation models and AI software for autonomy to enable machines to perceive, understand and navigate driving environment. Wayve’s first application is automated driving, developing a map-less and hardware-agnostic AI Driver for automakers and fleet owners to accelerate the path from assisted to automated driving on any vehicle.
AI Driver is designed to act as a driving intelligence layer for the automotive industry working on vehicles ranging from personally owned cars to robotaxis. It learns to drive from real-world data collected in more than 70 countries, which Wayve believes makes it a frontier training workload with physics and safety attached.
Seeing use in all levels of assisted and self-driving technology, the AI Driver software stack learns driving behaviour directly from large‑scale real‑world data, supporting adaptable performance across regions, road types and driving environments.
Wayve added that such models demand purpose-built cloud technology with reliability across long training runs and enough observability to prove one version is better than the last, and validate safety at scale.
As it looked to deliver a pre‑integrated advanced driver assistance and automated driving system for automakers, Wayve announced in March 2026 a technical collaboration with Qualcomm Technologies. This aims to see Wayve AI Driver deployed as an end‑to‑end driving intelligence layer for customers using Qualcomm’s Snapdragon Ride Platform and active safety software.
For its part, CoreWeave claims to be able to deliver industry-leading performance, demonstrated by record-breaking MLPerf benchmark results in inference and training.
Specifically, it noted its position as the only AI cloud to earn the top Platinum ranking in SemiAnalysis ClusterMAX three times in a row, and its top ranking for inference speed and price-performance for Moonshot AI’s Kimi K2.6 and Kimi K2.7 Code in independent inference benchmarking conducted by Artificial Analysis.
As part of its development plans, Wayve will now train its foundation models on CoreWeave Cloud with Nvidia Blackwell graphics processing units, expanding its AI infrastructure for training, evaluation and simulation. Wayve has already begun using weights and biases, and has now scaled into using CoreWeave Cloud.
“Our AI Driver gets better by running this data flywheel: gathering real-world driving experience, training on it, evaluating it and deploying only what outperforms what is already on the road,” said Silvius Rus, senior vice-president of engineering at Wayve. “Being able to access CoreWeave’s AI cloud, purpose-built for workloads like ours, helps accelerate the commercial deployment of our AI Driver.”
Gavin Jackson, vice-president of international at CoreWeave, said: “This agreement brings one of the most demanding physical AI programs in the world onto a platform purpose-built for it. As AI moves from the screen into the physical world, the companies leading that shift need training, evaluation and simulation infrastructure designed to work together from the start.”
