Wayve raises $200 million for learning-based autonomous driving
Wayve, a London-based developer of autonomous driving technology, has raised $200 million in funding. Since its founding in 2017, Wayve has raised more than $258 million.
The latest funding round was led by Eclipse Ventures. Microsoft also joined as an investor, as Wayve uses the Azure cloud computing platform to scale its machine learning technology.
Instead of relying on the traditional autonomous driving stack, HD maps and hand-coded rules, Wayve says it is building a data-driven, learnable driver that can scale, adapt and generalize its driving intelligence to new locations. Wayve applies a "camera-first" approach to its sensor suite and uses an end-to-end deep learning system. Like Tesla, Wayve does not use more expensive LiDAR sensors.
“As the industry struggles to solve self-driving with traditional robotics, it is becoming increasingly clear that AV2.0 is the right path to building scalable driving intelligence that can help commercial fleet operators deploy autonomy faster. Wayve is breaking new ground by building self-driving vehicles that can adapt to driving in new cities. As leaders in the field, they have assembled an exceptional team of machine learning experts and AV veterans to make AV2.0 a reality”
This is how Wayve describes the principle behind its autonomous driving technology: "We robustly model the world using computer vision, learning a representation of scene semantics, motion and geometry. We learn this representation together with learned planning and control algorithms. We produce a range of intermediate outputs with multi-task perception: semantic segmentation and instance segmentation, learned depth estimation, learned ego and scene motion, learned 3D detection, tracking and future prediction. We use surround-view cameras to perceive the world, using the latest advances in learning to learn how to plan motion for driving."
Wayve is working with Asda and Ocado, two leading grocery stores in the UK. A number of Ocado delivery vans equipped with Wayve's autonomous driving technology will be tested on urban delivery routes.
Wayve said the new funding will be used to expand its team, develop a Level 4+ prototype for passenger vehicles and delivery vans, scale deployment in partner fleets to launch last-mile delivery pilots, and build data infrastructure to strengthen its core autonomy platform.
“Today we have everything we need to take what we have built and push AV2.0 forward. We have assembled world-class strategic partners in transportation, grocery delivery and computing, as well as the best capital resources to scale our core autonomy platform, pilot products with our commercial fleet partners, and build the infrastructure for the global scaling of AV2.0”
The company's technology relies mainly on standard video cameras around the vehicle, tied to AI-driven automotive software, which makes it very adaptable to the environment because it depends less on a 4G or 5G network connection.
Alex Kendall said Wayve's test cars have successfully navigated not only London but also cities where their cars had never been before. Given that British streets typically have medieval layouts, this is no small feat.
“I think this funding round is a market signal that we are moving from validating the core technology to earning the right to scale and deploy commercially. When we started the business in 2017, it was the peak of the hype around autonomous vehicles, when billions of dollars had already been invested. Everyone thought it would happen within a year. Going off and building a contrarian startup to compete with trillion-dollar tech giants was perhaps a bit crazy. But milestones allowed us to move to the next level of scale. This is our ability to test in multiple cities. The fact that we trained the system in London and expanded across the UK, to Manchester, Coventry, Leeds and Liverpool. Plus all these commercial partners and the incredible talent we have been able to attract to our team”
"Imagine a world in which self-driving cars take us to work, pick our children up from school, bring us groceries and help our communities turn more parking lots into green parks. For this vision to move forward, autonomous driving technology must move forward and become more intelligent. It must be able to adapt to new cities, different landscapes and changing conditions. It has to cope with the unpredictable, the unknown and the never-before-seen. Like you, it must learn to react to any situation. Our AV2.0 technology completely rethinks how to solve self-driving by applying an approach based purely on machine learning," the company said.
