Uber and Wayve Win London's First Robotaxi Passenger Licenses

Photo by gogomoe on Unsplash

Uber and UK startup Wayve have secured the first London robotaxi license to carry paying passengers, beating Waymo and Baidu to the UK capital's paid market. Rides begin later this summer, and every vehicle will still carry a human safety driver in the front seat.

That last detail is the one to hold onto. The regulatory milestone is real, but the service launching in London is not driverless in the sense most riders expect from the word robotaxi.

Who Gets a Ride, and When

Invitations will go out from an existing Uber waitlist. According to reporting on the license approval, selected passengers from a pool of 100,000 Londoners already registered on Uber's interest list will receive booking access later this summer of 2026, with rides delivered through the standard Uber app rather than a separate Wayve product.

The fleet operates under private hire rules, not black cab rules. Approval conditions reference London's Vision Zero goal of eliminating all road deaths and serious injuries by 2041, which is why Transport for London tied the license to a supervised operating model rather than a leap straight to empty driver seats.

Expect an early Wayve robotaxi launch date measured in weeks, not months. The Uber-Wayve announcement stops short of naming a specific first-ride day, which is consistent with a regulator that wants operators to control ramp pace rather than pre-commit to a headline date.

The Safety Driver Clause Is the Whole Story

Read the license carefully and one condition dominates everything else. A licensed private hire driver will occupy the driver seat during every trip, oversee the vehicle, and take manual control if needed. This is not a driverless commercial deployment. It is a supervised autonomy pilot with paying passengers.

That framing matters for anyone comparing this rollout to Waymo's Phoenix or San Francisco operations, where riders sit in cars with no one up front. London's paid service will look, from the back seat, very similar to a normal Uber. The car drives itself; a human is contractually and physically responsible.

The economics of that model are unforgiving. Wayve and Uber still pay a driver per trip, so the unit cost of a London autonomous taxi passengers ride is closer to a standard private hire fare than to the labor-free economics Waymo has been chasing. The license is a foothold, not a profit engine. It buys real-world route data, regulator trust, and a public brand presence while the harder question of removing the safety operator gets negotiated separately.

That is the trade the two companies made. Speed to market in exchange for keeping a person in the loop, in a city where a fatal incident would set the entire UK program back years.

Wayve Uber Self-Driving London Puts Waymo and Baidu on the Back Foot

The competitive picture in the UK capital has now flipped. Alphabet's Waymo and Chinese firm Baidu continue testing autonomous fleets in London under non-passenger trial arrangements only, which means they can drive on public roads but cannot charge riders. Uber and Wayve can.

Waymo's London plans are not modest. Waymo's London fleet will use all-electric Jaguar I-Pace vehicles equipped with the Waymo Driver system, with fleet operations and charging handled by Moove. That is a serious commitment of hardware, but a testing permit and a passenger license are different regulatory objects, and Wayve is currently on the right side of the line.

Baidu's position is stranger still. Lyft and Baidu announced plans to launch Baidu's RT6 robotaxis on the Lyft app in Germany and the UK in 2026 pending regulatory approval, and those approvals had not been granted. In the UK specifically, Lyft has no operational network, so Baidu's route to London riders runs through a partner that itself has no local rider base. Wayve, by contrast, is British, London-based, and already integrated with the country's dominant ride-hail app.

Scale context helps here. Apollo Go, Baidu's robotaxi arm, has cumulatively provided 20 million rides worldwide and accumulated 300 million kilometres of autonomous driving. Wayve has nothing close to that operational history. What Wayve has instead is a passenger license in a G7 capital, which is the exact document Baidu is missing.

The Europe Robotaxi 2026 Map Just Redrew Itself

London is no longer the first European city with paying robotaxi riders. That title belongs to Zagreb. Croatia launched Europe's first commercial robotaxi service in April 2026, operated by startup Verne using Pony.ai's seventh-generation autonomous driving system.

Marko Pejković, Verne's Co-Founder and CEO, described the launch as "the first time in Europe there is a real commercial robotaxi service" where members of the public can take genuine autonomous rides. Verne's Zagreb vehicles also carry trained safety operators onboard during the early deployment phase, with fully driverless operation described as a future target pending further regulatory work.

That means both of Europe's first paid robotaxi services, Zagreb and London, ship with a human in the vehicle. The story being sold in press releases is autonomy. The story on the ground is supervised autonomy at commercial scale, which is a meaningfully different thing.

Croatia got there first partly because it changed its rules first. Croatia updated its legal framework to enable driverless licences ahead of most EU member states, giving a country of under four million people a genuine first-mover advantage on a technology usually associated with San Francisco and Shenzhen. The lesson for policymakers in Paris, Rome, and Amsterdam is direct: the legislation is the product.

Other capitals are moving. London, Luxembourg, Madrid, and Munich are actively competing for 2026 robotaxi deployments alongside Zagreb, with Bolt, Uber, Wayve, Waymo, WeRide, and Pony.ai all lined up across European cities. Croatia and Luxembourg have already updated their legal systems, Spain is consulting on a draft bill, and Germany made Traffic Act changes in 2021 enabling pilot deployments. The rest of the EU has been slower, which limits how quickly the London model can be exported.

Uber's European Strategy Is Not About Uber Building Cars

Uber does not need to build a self-driving stack. It needs a passenger app on top of one, and it is aggregating partners fast enough that no single failure kills the strategy.

The London deal with Wayve sits alongside a parallel push in Switzerland. WeRide and Uber announced on June 17, 2026 a plan to launch a commercial robotaxi service in Zurich, described as WeRide's first deployment region in Switzerland and subject to regulatory approval. The Zurich announcement was Uber's second European announcement with WeRide, indicating an earlier market agreement already in place.

WeRide uses an asset-light operating strategy with Rydera as fleet operator, which lets Uber plug into a local fleet manager rather than owning or maintaining vehicles itself. That is the pattern to watch. Uber pairs an AV developer with a regional fleet operator, keeps the demand-side app, and lets the two suppliers absorb the capital cost of the vehicles.

For readers tracking the wider Auto Tech articles beat, this is closer to how airlines negotiate slot rights than how car companies sell cars. The scarce asset is regulatory permission, not the vehicle.

The Revenue Reality Behind the Headlines

Robotaxi licenses generate press. Robotaxis, so far, do not generate much revenue. Pony.ai reported Q4 2025 robotaxi revenue of $6.7 million, a 159.5% year-over-year jump from $2.6 million in Q4 2024, which sounds like rapid growth until you notice the base is small enough to round to zero at the scale of a global auto business.

That context should temper interpretations of the London announcement. A city-scale license is a serious operational and political achievement, and it can still be true that the underlying business does not yet clear the cost of the safety drivers in the front seats. The industry is at the stage where regulatory wins outpace unit economics, which is a normal position for a hardware-heavy technology but a dangerous one to confuse with commercial maturity.

WeRide's non-European credentials illustrate what a further-along regulatory position looks like. WeRide holds a landmark Abu Dhabi commercial license described as the world's first urban Level 4 commercial license granted outside the United States. London's approval is a step behind that on the autonomy ladder, because it still requires an onboard human, but a step ahead on brand exposure given the size of the London ride-hail market.

What Actually Runs on London Streets Next

The Wayve fleet will operate on Uber's app under the safety-driver-in-seat model until Transport for London approves a variation. There is no published timeline for removing the driver.

Related European pilots are landing on similar structures. In Berlin, BVG and MOIA are launching autonomous VW ID. Buzz AD shuttles in the NoWeL4 project, with five vehicles operating in a 15 sq km area serving roughly 80 stops, and passenger trials with safety operators started in the first half of 2026. WeRide's Robobus in Leuven, Belgium holds that country's first federal Level 4 test permit, running as an autonomous shuttle between the main station and a nearby corridor with a safety officer present.

Every one of these deployments keeps a human in the vehicle. That is the honest state of European robotaxi service in 2026.

For context on the regulatory instruments that will eventually approve or block a driverless step, our guide to how to look up and interpret an NHTSA robotaxi software recall explains the US disclosure system that UK regulators will study when writing their own reporting rules. The UK does not yet have an equivalent public database, and how that gap gets closed will shape whether London can scale beyond a single operator.

The Specific Thing to Watch

Watch for the first Transport for London variation notice that removes the safety driver from a defined Wayve operating zone. That document, not a press release, is what turns the current London robotaxi license into an actual driverless service, and it will be the moment when the economics change enough to matter to Uber's margins and to competing operators from Waymo to Baidu.

Until that notice appears, London has a paid autonomous taxi service in name and a supervised ride-hail service in practice. The distinction is worth remembering the first time an invited passenger books a car.