Wayve vs. Tesla and Waymo: Why Licensing AI Scales Faster

Wayve approaches self-driving differently from Tesla or Waymo. Instead of hand-coding rules for every situation on the road, the company uses end-to-end learning: a single neural network takes in camera feeds and outputs driving commands directly. It is a fundamentally different approach, and CEO Alex Kendall explained in an interview with Bloomberg why he believes it is the future of the industry.

One algorithm instead of a modular architecture

Tesla and Waymo split the autopilot into separate modules. One recognizes objects on the road, a second predicts their movement, a third plans the route, and a fourth controls the throttle, brakes and steering. Each module is trained separately, and the results are passed down the chain to the next one.

Wayve calls this archaic. Its approach is closer to how the human brain works: you look at the road and immediately know what to do with the wheel. There are no intermediate steps, no data handoffs between modules, no points of failure. This simplifies the system and speeds up adaptation to new conditions. If road markings in London differ sharply from those in San Francisco, you don't need to retrain three separate modules; it is enough to refine one algorithm.

"The modular approach looks better on paper, but end-to-end works better in reality,"

Kendall said in the interview.

Licensing as a fast path to scale

Tesla builds its own cars and embeds the autopilot in the hardware. Waymo builds its own robotaxi fleet. Both pour huge sums into manufacturing, logistics and maintenance.

Wayve chose a third path: it licenses its AI algorithm to automakers. Partners build the code into their cars, and that's it. Wayve doesn't need to invest billions in producing vehicles, delivering them and maintaining a fleet. It simply sells a software license, and partners fund development through their payments.

• Minimal capital investment to reach scale
• Partners themselves fund the technology's development
• One algorithm runs on millions of cars across different brands
• Wayve focuses only on AI, not on hardware and logistics

What a small startup is betting on

Competition is fierce. Tesla has already equipped millions of cars with cameras and collects telemetry every day. Waymo is backed by Google and has years of testing on American roads. Next to these giants, Wayve is a small London startup.

But Kendall isn't panicking. He is betting that the right algorithm and the right business model beat scale and corporate weight. If AI licensing becomes the industry standard, Wayve could overtake both.

What this means for the industry

The future of autopilots may not be a fight among a few titans at all. It may be an ecosystem in which one or two top AI algorithms are licensed to hundreds of partners. If Wayve has calculated correctly, the industry flips: instead of a race to build cars, a race for algorithms and licensing contracts.