The autonomous driving industry faces significant challenges because traditional methods rely on high-definition maps, location-specific engineering, and rule-based systems. This has slowed down widespread commercial use.
Wayve, a London-based leader in embodied AI for autonomous driving, takes a fresh approach with its end-to-end AI platform. Rather than using maps or fixed rules, Wayve’s AI Driver learns from a wide range of real-world driving data to see, understand, and navigate complex environments.
The system runs only on onboard computing and native sensors, so it can be used in new cities without any special adjustments. In 2025, Wayve became the first autonomous vehicle company to achieve zero-shot driving in over 500 cities across Europe, North America, and Japan.
To move from research to real roads, Wayve recently raised a massive $1.5 billion, pushing its valuation to $8.6 billion. The $1.2 billion Series D round was led by Eclipse, Balderton, and SoftBank Vision Fund 2, with new backing from Ontario Teachers’ Pension Plan, Baillie Gifford, British Business Bank, Icehouse Ventures, and Schroders Capital.
Strategically, Microsoft, NVIDIA, and Uber are involved, along with car giants like Mercedes-Benz, Nissan, and Stellantis. Uber is also providing additional milestone-based funding to help launch Wayve-powered robotaxis in more than 10 global markets.
A vehicle-agnostic AI platform that brings autonomy to everyone
Wayve’s story started in 2017 with Alex Kendall, a Cambridge PhD who saw that rule-based autonomous vehicles wouldn’t scale. So he aimed to build embodied AI that learns to drive from data, much as large language models learn to write.
Wayve’s end-to-end AI is trained on driving data from more than 70 countries. The system processes raw sensor input and converts it into safe driving actions, skipping the usual perception and planning steps — and yes, those HD maps.
It is already deployed in over 500 cities (219 without any local data) without fine-tuning, unlike competitors who require extensive per-city training. It runs on embedded computing and native sensors across brands, with automakers customising through Wayve’s toolchains.
A single model covers everything from supervised assistance to driverless operation, continuously improving through fleet data.
Unlike Waymo, Cruise, Tesla FSD, or Zoox, Wayve stands out with its mapless, generalised approach. It offers a licensable software model rather than Tesla’s vertically integrated system.
It also works with multiple OEMs, such as Mercedes, Nissan, and Stellantis, rather than focusing on just one brand. Its diverse data from 70 countries and zero-shot capability give it an advantage for global scaling without localisation costs.
What’s next?
With $1.5 billion in funding, Wayve is preparing for a large-scale commercial rollout. In 2026, Uber’s robotaxi trials will start in London, with plans to expand to more than 10 global markets using L4-ready vehicles from partner OEMs.
In 2027, you can expect to see Wayve’s AI Driver in consumer cars, beginning with Nissan’s L2+ hands-off assistance and then advancing to L3 and L4 levels.
