Skip to main content

Wayve signs multiyear CoreWeave deal for self-driving AI training

The London startup will use Nvidia Blackwell chips for its AI Driver, which learns from driving data collected in more than 70 countries.

By Mei Ling Tan

Illustration of a car connected to cloud computing infrastructure
Photo: TNW (Financial Times)

CoreWeave said London-based self-driving software startup Wayve will use its cloud infrastructure under a multiyear agreement to develop its AI Driver. The companies did not disclose the contract’s value or its precise duration.

The work will run on Nvidia Blackwell chips and cover model training, testing and simulation. Wayve’s software learns from driving data collected across more than 70 countries. Silvius Rus, Wayve’s senior vice president of engineering, said the company evaluates models trained on real-world driving experience and deploys them only when they outperform the software already in use.

Rather than manufacture vehicles, Wayve licenses software to automakers and ride-hailing companies. Uber began taking passengers in London in September using Wayve-driven vehicles, each accompanied by a safety driver. Mercedes-Benz plans to incorporate the AI Driver into production vehicles.

Wayve has also trained models using Microsoft’s Azure cloud since 2022. Its investors include both Microsoft and Nvidia. The startup previously used the model-tracking tool Weights & Biases, which CoreWeave acquired.

Keep reading: Tech News

All Tech News stories

Latest in Technology

All latest
  1. Gadgets

    Huawei claims all-Chinese components for Qinyun HM650 desktops

  2. Tech News

    Disney names Adobe as AI partner for models based on its characters

  3. Tech News

    Meta expands AI age checks for Instagram teen accounts worldwide

  4. Tech News

    Nvidia unveils GR00T 2 robot model trained partly on human video

  5. Tech News

    China expands Apple Intelligence registration to iPads and Macs

  6. Tech News

    Google adds TPU variants for confidential computing and long-context AI