British artificial intelligence company Wayve has signed a definitive production agreement with Mercedes-Benz that will put its AI driving system into future production vehicles, moving the London-based company further from development trials towards commercial-scale deployment.
The companies expect the first customer applications to begin within the next two years. The agreement covers advanced point-to-point driving assistance for urban roads and highways and follows a multi-year technical collaboration between the businesses.
Mercedes-Benz also invested in Wayve’s Series D financing earlier this year. The new agreement establishes a production route for technology that the companies have already integrated with Mercedes-Benz vehicle hardware, the MB.OS operating system, and its mapping interfaces.
The arrangement does not yet identify which Mercedes-Benz models will carry the system, the markets in which it will initially be offered, or the regulatory approvals required. Those details will determine how quickly a production agreement becomes a widely available customer feature.
Wayve describes its approach as “AV2.0”, using end-to-end artificial intelligence to learn driving behaviour rather than relying primarily on detailed, manually engineered rules and high-definition maps. Its models are trained on Nvidia computing infrastructure hosted on Microsoft Azure, while the company says the same underlying AI architecture can be developed across different levels of driving automation.
Alex Kendall, Wayve’s co-founder and chief executive, said: “The pace and depth of our development work together reflect a deep technical alignment between our teams.”
Automotive AI companies ultimately need production relationships as well as successful demonstrations. Testing can establish technical capability, but deployment within a global carmaker introduces additional requirements around safety engineering, vehicle integration, manufacturing cycles, software maintenance, liability, cybersecurity, and regulatory approval.
The agreement gives Wayve access to a premium vehicle platform as established manufacturers increasingly combine internal software development with specialist AI partnerships. Automated-driving systems require large amounts of computing power and training data, while rapid changes in AI development place pressure on traditional vehicle programmes that can run for several years.
Wayve’s commercial activity has accelerated during 2026. Earlier in September, the company and Uber launched supervised autonomous rides in London, creating another route for its technology to operate in real-world passenger services. The Mercedes programme is structurally different because it is designed around integration into consumer vehicles rather than a ride-hailing fleet.
That distinction broadens the potential market for Wayve’s software. Technology that can be integrated across multiple vehicle architectures would allow the company to sell into manufacturers without carrying the capital cost of producing vehicles itself. Carmakers, meanwhile, can gain access to specialist AI capability without developing every element of an automated-driving stack internally.
Commercial models across autonomous driving have also become more varied. Early industry expectations centred heavily on purpose-built robotaxi fleets, but manufacturers are now pursuing approaches ranging from increasingly capable driver-assistance systems to highly automated vehicles intended to operate without human control under defined conditions.
Wayve says its underlying AI Driver is designed to scale across that range. The Mercedes agreement does not mean the first vehicles will operate without a driver. The announcement centres on advanced driving assistance, and the exact boundaries of the system will depend on the final specification and approvals in individual markets.
Regulation remains a significant variable because automated-driving rules differ between jurisdictions. Functions that can be activated in one country may require additional technical evidence, operational restrictions, or driver-monitoring systems in another, potentially creating different launch timetables from the same underlying vehicle platform.
Vehicle manufacturers must also manage software throughout the useful life of a car. Systems based on rapidly developing AI models will require processes for updates, validation, incident analysis, and compatibility with hardware already on the road. That shifts part of automotive product management towards a continuing software relationship rather than a fixed specification at the point of manufacture.
Wayve and Mercedes have established the production framework and completed integration work with the manufacturer’s architecture. Vehicle selection, regulatory clearance, customer pricing, and the eventual operational limits of the system remain to be disclosed as the first launch approaches.
You must be logged in to post a comment.