Volkswagen has picked the British AI company Wayve to supply the driving software for its future cars, according to Handelsblatt, which cites three people familiar with the decision. The company it beat is Nvidia — which is also one of Wayve's investors.
Neither Volkswagen nor Wayve has announced anything. Treat the whole of this as reporting until one of them does.
What is reported to have happened
Handelsblatt describes a contest. In August, Volkswagen is said to have gathered Wayve, Nvidia and around a dozen other autonomous-driving software suppliers in a southern German city and had them demonstrate against each other. Wayve's bid is reported to have convinced on both technical and commercial grounds.
The software would go to Cariad, Volkswagen's in-house software unit, which would integrate it into a stack the group has already built — its own choice of chips, sensors and surrounding software.
That is the structural point, and it is why this is not simply "VW picked a supplier".
| Option | What Volkswagen would have bought |
|---|---|
| Nvidia | A complete autonomous-driving system — compute, sensors and software together |
| Wayve (reported choice) | The AI driving model only, integrated by Cariad into Volkswagen's own stack |
Volkswagen is reported to be negotiating a separate chip deal with Nvidia regardless. So the outcome is not that Nvidia lost Volkswagen; it is that Volkswagen declined to buy autonomy as a finished product and kept the integration work in-house.
Why a Tesla owner should care
Volkswagen is the largest carmaker in Europe. Whatever driver-assistance system it ships is the system most European drivers will experience, and the one FSD will be compared against in showrooms here long before any robotaxi argument is settled.
The architectural choice is the interesting part. Tesla builds the entire stack itself — its own inference chip, its own cameras, its own model, trained on its own fleet — and that vertical integration is the thing Tesla says makes FSD possible. The reported Volkswagen approach splits the difference: buy the hardest piece, the AI driver, and own the integration.
Wayve is already doing this for other people. It has a production agreement with Mercedes-Benz to put its AI Driver in cars within two years, a partnership with Stellantis announced in May, and a deal with Nissan to supply the next generation of ProPILOT from 2027. Nissan, Mercedes-Benz and Stellantis are all also investors in the company, alongside Nvidia, Microsoft and Uber. Adding Volkswagen would make Wayve the default driving intelligence for a large part of the European and Japanese industry.
For Tesla that is a different competitive shape than a dozen carmakers each building their own system badly. One good supplier spread across many brands closes the gap faster than any of them would alone.
The other Volkswagen autonomy project is not this one
It is worth keeping the two apart. Volkswagen's robotaxi arm, MOIA, runs the self-driving ID. Buzz on a separate technical track, and it is in trouble: its first round of investor talks collapsed in September with both Uber and Lyft walking away, after roughly 2 billion euros spent.
A Wayve agreement would address consumer cars, not that. If anything it reads as the group deciding that the assisted-driving feature it sells to private buyers is the part worth winning, and buying the best model available rather than waiting for its own.
Nothing here is confirmed, no contract value has been reported, and no timing has been given for when a Wayve-powered Volkswagen would reach customers.