Wayve Raises $1.5 Billion at $8.6 Billion Valuation as BMW and Mercedes Abandon Level 3, Signaling a Tectonic Shift in Autonomous Driving
UK startup Wayve secures $1.5B from NVIDIA, Microsoft, Uber, and three automakers for its AI foundation model approach to self-driving, just as BMW and Mercedes shelve their in-house Level 3 programs.
Overview
The autonomous driving industry crossed a strategic inflection point this week. London-based startup Wayve closed a $1.2 billion Series D round — part of a broader $1.5 billion capital package — valuing the company at $8.6 billion, according to Wayve’s official announcement. The round drew commitments from NVIDIA, Microsoft, Uber, Mercedes-Benz, Nissan, and Stellantis, making it one of the largest-ever raises for a European AI company.
The timing is significant. Days before the Wayve announcement, BMW confirmed it would remove its Level 3 “Personal Pilot” system from the 7 Series, following Mercedes-Benz’s January decision to shelve its own “Drive Pilot” Level 3 technology from the S-Class. Two of the world’s most prestigious automakers have now concluded that building autonomous driving in-house is too costly and too limited in scope — and at least one of them, Mercedes, is putting capital into Wayve instead.
What We Know
Wayve’s Series D was led by Eclipse, Balderton, and SoftBank Vision Fund 2, with additional institutional investors including Ontario Teachers’ Pension Plan, Baillie Gifford, and the British Business Bank, as reported by Tech.eu. Beyond the $1.2 billion core round, Uber committed additional milestone-based capital to fund multi-year robotaxi deployments, bringing the total secured capital to $1.5 billion, according to Wayve.
The company’s pitch centers on a fundamentally different architecture than the industry’s dominant player, Waymo. Where Waymo relies on a sensor fusion stack augmented with high-definition maps and city-specific engineering, Wayve trains a single end-to-end foundation model that processes raw sensor data and produces driving decisions directly. Crucially, the system runs entirely on onboard compute and does not require HD maps, as detailed by CleanTechnica.
Wayve claims to be the first autonomous vehicle developer to drive “zero-shot” in more than 500 cities across Europe, North America, and Japan — meaning its system operated without any city-specific fine-tuning before deployment, according to the company. This generalization capability stems from a foundation model trained on driving data from over 70 countries and multiple vehicle platforms.
The commercial roadmap is aggressive. Wayve and Uber plan to launch robotaxi trials in London in 2026, with expansion to 10 or more global markets. Starting in 2027, automakers will begin selling consumer vehicles equipped with Wayve’s AI Driver, beginning with Level 2+ “hands-off” capability and progressing toward Level 3 and Level 4 “eyes-off” autonomy, as reported by TechCrunch.
“Autonomy will scale through a trusted platform that automakers and fleets can deploy globally and improve continuously,” CEO Alex Kendall said in Wayve’s announcement. Microsoft CEO Satya Nadella called Wayve’s work the “frontier of embodied AI for autonomous driving.”
The Automaker Retreat
The automaker investment in Wayve arrives against a backdrop of retreat from proprietary autonomy programs. BMW’s “Personal Pilot L3” system, introduced in 2024 and priced at 6,000 euros, will be discontinued with the 7 Series facelift in late April 2026, according to Electrive. The system required expensive LiDAR sensors and only functioned on motorways at speeds below 60 km/h — a narrow use case that failed to attract meaningful buyer interest.
BMW will replace it with a “Motorway Assistant” Level 2 system priced at 1,450 euros that allows hands-off steering at up to 130 km/h but requires driver attention, Electrive reported. Mercedes-Benz similarly pivoted to a Level 2++ system called “MB.Drive Assist Pro” after removing Drive Pilot from its refreshed S-Class.
The pattern is clear: the liability, regulatory complexity, and limited operational domain of Level 3 — where the car assumes legal responsibility for driving — proved more burdensome than the engineering challenge itself.
The Competitive Landscape
Wayve’s rise does not happen in a vacuum. Waymo, backed by Alphabet, expanded its robotaxi service to four more U.S. cities this week — Houston, Dallas, San Antonio, and Orlando — bringing its total to 10 American metros and putting it on pace to serve over one million rides per week by year-end, according to CNBC. Waymo raised $5.6 billion last year and remains the operational leader by a wide margin.
But the two companies represent different theories of how autonomous driving will scale. Waymo operates its own fleet with custom-built sensor suites and city-by-city expansion. Wayve sells software to automakers and fleet operators, arguing that a generalizable AI model can reach more markets faster and at lower cost. Uber CEO Dara Khosrowshahi endorsed this view, calling Wayve’s end-to-end approach “purpose-built for scale, safety, and effectiveness,” per Wayve.
What We Don’t Know
Wayve has not disclosed detailed safety metrics or disengagement rates from its zero-shot city deployments. The claim of operating in 500-plus cities without fine-tuning is striking, but the conditions of those drives — whether they were supervised, limited in scope, or on specific road types — remain unclear.
The London robotaxi trial timeline is ambitious. Regulatory approval for commercial autonomous vehicle operations in the UK has not yet been granted, and the timeline for that process is uncertain.
Whether automakers’ investments in Wayve represent genuine strategic commitment or hedging behavior also remains to be seen. Mercedes-Benz, Nissan, and Stellantis are all investors, but none have announced exclusive partnerships or firm volume commitments for consumer vehicles.
Finally, the question of whether end-to-end AI systems can match the safety record Waymo has established with its more traditional stack remains the central technical debate in the field. Wayve’s foundation model approach is theoretically more scalable, but it has yet to demonstrate the millions of uncrewed miles that Waymo has accumulated.