The startup is betting on end-to-end AI to power the next generation of autonomous vehicles.
London-based autonomous driving startup Wayve is gaining momentum as investors and automakers increasingly back its AI-driven approach to self-driving technology. The company has raised $2.8 billion from a group of investors and strategic partners, including Nvidia, Mercedes-Benz and Nissan, while recently announcing plans to deploy its software in robotaxis built by Stellantis for Uber’s ride-hailing platform.
Founded in 2017 by CEO Alex Kendall, Wayve is developing an end-to-end machine learning system that allows vehicles to interpret sensor data and make driving decisions without relying heavily on pre-programmed rules or detailed high-definition maps. Instead, the AI learns from real-world driving data, enabling vehicles to adapt to unfamiliar environments much like a human driver.
Unlike Tesla, which primarily depends on cameras, Wayve’s platform is designed to work with multiple sensor types and AI chips, making it easier for different automakers to integrate the technology into their vehicles. The company aims to license its software across brands rather than build its own fleet.
Interest in autonomous driving has been reignited by the commercial expansion of Alphabet’s Waymo, whose robotaxi services now operate in multiple cities. At the same time, more developers are adopting end-to-end AI models as computing power and AI capabilities continue to improve.
However, the technology remains a subject of debate. Critics argue that AI-driven systems function as “black boxes,” making it difficult to understand how vehicles arrive at specific driving decisions. Waymo continues to combine end-to-end AI with traditional rule-based software and mapping, saying that hybrid systems remain essential for ensuring safety at scale.
Wayve maintains that its AI-generated safety models allow vehicles to respond more effectively to unpredictable road conditions than rigid, pre-programmed systems. Even so, industry experts say both AI-first and conventional approaches have strengths, and widespread deployment of fully autonomous vehicles will still require years of testing, validation and regulatory approval.
