London's robotaxi dream pushed to 2025 as regulators and engineers hit the brakes

Uber and Wayve had licences in hand and a 'later this summer' promise on the table. Now, people waiting for a driverless cab in London will have to wait a lot longer.

AI2Day NewsdeskAI-assistedPublished Updated Editor: Lee Brown3 min read
Illustration: A large modern warehouse interior
Illustration made with AI. Not a photograph of the events described.
Share

Key points

  • Uber and Wayve received the first minicab licences in London allowing self-driving taxi rides to paying customers, but with a human safety driver present for now.
  • Both companies had publicly promised to begin passenger trips "later this summer" before a full driverless rollout.
  • Regulatory gaps and unresolved technical requirements have pushed the London robotaxi launch beyond 2024.
  • The UK government had positioned London as an early showcase for autonomous transport on public roads.

A robotaxi drives itself, using onboard AI software and cameras rather than a human at the wheel. London was supposed to get its first taste of them this summer. It will not.

As reported by The Guardian, Uber and the London-based self-driving company Wayve were granted the capital's first minicab licences covering autonomous passenger rides. That was a genuine milestone. The catch: a trained human safety driver still sits behind the wheel. Full driverless operation is, in theory, coming later.

We covered Wayve's progress in our 26 August 2026 story on Transport for London's role in this rollout. This delay shows how quickly a promising timeline can unravel.

Why is this taking longer than planned?

Regulators haven't published clear rules telling companies what their vehicles must do to qualify for unsupervised operation on London streets. Without those rules, companies can't certify their technology meets the standard, and without certification, no paying rides without a safety driver.

Technical hurdles are real too. Wayve's system uses what the industry calls end-to-end AI: a single neural network (a type of AI loosely modelled on how the brain processes information) watches the road and decides what the car does next. It's a newer, less road-tested approach than the hardware-heavy systems used by rivals like Waymo in the United States. Proving it reliable enough for regulators takes time.

The UK government had hoped the country would become a visible leader in autonomous transport. That ambition is now at least a year behind schedule.

What does this mean for ordinary Londoners?

For now, nothing changes day-to-day. If you catch an Uber in London and it happens to be a Wayve-equipped vehicle, a human driver is still in control. You're not a test subject.

When commercial robotaxi rides do eventually launch, the experience will likely start cautiously: fixed routes, limited hours, a safety driver on board. Think of it less like hailing a cab and more like a supervised trial that happens to charge a fare.

Cost is worth watching. Human drivers account for the bulk of a taxi fare, and removing that cost could lower prices. But the computers and sensors running a self-driving car currently add tens of thousands of pounds per vehicle. That bill has to be recovered somewhere.

What happens next?

Wayve and Uber haven't pulled out. Both licences remain valid, and the companies are understood to be working with Transport for London and national regulators toward a revised timeline. A 2025 start for limited paid rides is the realistic target now.

Publishing clear technical standards is the single biggest step regulators could take. Once companies know exactly what bar they must clear, engineering teams have a concrete target to hit. Until then, the gap between a licence and an actual driverless passenger isn't a legal formality. It's the whole problem.

© 2026 AI2Day