Wayve CEO Says AI-Driven Self-Driving Approach Built on ‘Fraction’ of Tesla’s Data

Wayve CEO Alex Kendall has defended his company’s artificial intelligence-first approach to autonomous driving, arguing that a foundation model capable of learning from real-world driving can scale more efficiently than competing systems from Waymo, Tesla, and Chinese rivals like Baidu and Pony AI.

In an interview, Kendall outlined three emerging business models in the autonomous vehicle industry: selling own-brand cars (Tesla’s model), building proprietary robotaxi fleets (Waymo’s approach), and licensing technology to third-party manufacturers and fleets — the path Wayve has chosen.

“I’m sure these three business models will co-exist,” Kendall said. “But we’ve chosen a business model that we believe is going to be the largest because the majority of fleets or manufacturers of robots … are going to find a more efficient and more effective way to partner.”

Kendall claimed that Wayve has achieved safety performance comparable to Tesla’s “with a fraction of the data and compute” used by that solution. He added that scaling up with data from global consumer fleets will further improve Wayve’s system.

When asked about Waymo’s argument that lidar and radar create redundancies that improve safety, Kendall called it “a complex and nuanced question” that depends on the product. “We believe that different sensor configurations are going to be beneficial for different products,” he said, noting that Wayve’s foundation model can operate with camera-only systems, radar, or lidar depending on the vehicle.

A live demonstration of a Wayve-powered car in London’s Kings Cross area operated without a safety driver touching the wheel. The vehicle used six cameras and one radar — a sensor stack Kendall said costs “hundreds of dollars” and is suitable for mass-market vehicles.

“We have an AI on board the vehicle that’s currently driving this car,” Kendall said during the ride. “The computer’s actually on the device. All decisions made on board using onboard intelligence.”

Kendall contrasted Wayve’s end-to-end learning approach with traditional autonomy companies that build separate detectors for indicators, cars, and traffic lights with programmed logic. “We don’t tell the car how to behave,” he said. “We simply say, ‘Hey, here’s the end outcome you need. Let the data speak for itself.’”

The CEO also highlighted London’s unique challenges — including roughly 20 times more roadworks and ten times more cyclists and pedestrians than San Francisco — as an advantage that forced Wayve to develop a more scalable approach. “Starting in London, we were out of the thought bubble of Silicon Valley … to build something contrarian.”

Kendall, a New Zealander running a global company from London, acknowledged feeling pressure but said he wants “to break out and build a global success story.” He added: “This is an adventure. This is the chance to push the frontier, both technically and now commercially.”