US ride-hailing companies Lyft and Uber are partnering with Chinese self-driving tech giant Baidu to trial driverless taxis in the UK next year.

The plans could see hundreds of robotaxis take to the roads in London

The plans, which are pending regulatory approval, could see hundreds of robotaxis take to the roads in London, following work by Waymo to debut autonomous vehicles (AVs) in the capital. London-based start-up Wayve is also preparing ​to launch driverless trials in 2026.

Lyft chief executive David Risher announced the plans on X and said riders across the city would be the first in the region to experience Baidu’s Apollo Go vehicles.

“We expect to start testing our initial fleet with dozens of vehicles next year – pending regulatory approval – through the Lyft and Freenow ecosystem, with plans to scale to hundreds from there,” he explained.

Risher elaborated: “Our hybrid network – AVs and drivers working together – will serve all of London’s travel needs. We’re working closely with Transport for London regulators and local communities to expand the rideshare market and create more opportunities for everyone. More to come, but can’t wait to see the future of mobility arrive in one of the world’s greatest cities.”

Transport Secretary Heidi Alexander said it was “another vote of confidence in our plans for self-driving vehicles”. The news follows the Government’s call for evidence on developing the regulatory framework for autonomous vehicles.

Alexander continued: “We’re planning for self-driving cars to carry passengers for the first time from spring, under our pilot scheme – harnessing this technology safely and responsibly to transform travel.”

Meanwhile, Uber said its first pilot in the UK is expected to start in the first half of 2026.

“We’re excited to accelerate Britain’s leadership in the future of mobility, bringing another safe and reliable travel option to Londoners next year,” it said.

Jonny Berry, head of decarbonisation, innovation & strategy at Novuna Vehicle Solutions, said the UK is progressing steadily with infrastructure, regulation and data readiness for autonomous vehicles but warned that scaling nationally by 2030 will require addressing several known limitations – including “inconsistent data quality, rural infrastructure gaps, lack of system interoperability and a shortage of trained operational staff”.

He also said that public trust and real-world operational learnings would “also play a crucial role in determining the pace and scope of rollout”.

Berry also gave some insights on the practical challenges of deploying autonomous vehicles on UK roads – from infrastructure and data readiness to public trust and operational viability.

“Typically, narrower UK roads, especially in rural or older urban areas leave less margin for autonomous vehicles to manoeuvre, requiring ultra-precise path planning and perception systems.

“Distinctive road signage, unique lane markings and signal sequences in the UK mean AV systems trained in the US must be recalibrated. Roundabouts are far more prevalent in the UK than the US, which uses more traditional signal-controlled intersections – a challenge for AV decision-making.

“Rain, fog and reduced daylight during winter in the UK impact sensor performance and require advanced fusion of radar, lidar and cameras. High pedestrian and cyclist density, prevalent on UK roads, necessitates AV’s needing advanced behaviour prediction to safely coexist with vulnerable road users in urban areas.

“Crucially, AVs must localise behaviour to UK road conventions and adapt models to avoid unsafe assumptions.”

Berry also warned of EV infrastructure differences and the implications for autonomous fleets.

“Autonomous fleets require high charger density within defined operating domains. There are currently over 80,000 charge points in the UK, with urban concentration and rural gaps. Sparse coverage in rural areas limits the Operational Design Domain (ODD) of AVs and restricts where they can be deployed.

“Even in urban areas where rapid charging access is growing fast in the UK, AVs cannot queue or reroute easily. Most public charging sites are unmanned and require manual connection, which AVs cannot currently do without human intervention. This limits uptime and reduces the practicality of autonomous public charging unless solutions like robotic charging or attended hubs are introduced.

“Autonomous fleets depend on consistent charger reliability to remain in service. Charger faults must be rare and remotely diagnosable for AV viability.”

Berry also warned that the UK’s HD mapping coverage is strong in cities but remains inconsistent in rural areas, which limits AV route flexibility and scale.

“The UK has developed a strong testbed network (e.g. CAM Testbed UK), but large-scale autonomous trials on public roads have not progressed beyond early-stage research and simulation. While the physical and digital infrastructure exists for controlled testing, the absence of real-world commercial AV trials limits learnings on edge cases, operational readiness and public interaction.

“AVs are highly connected systems, making them potential targets for cyber-attacks and data breaches. The UK enforces UNECE R155 cybersecurity standards and strict GDPR compliance.

“Autonomous fleets must demonstrate secure communications, protected user data, and resilience against unauthorised access, tampering or remote interference. Failure to do so risks loss of public trust, operational bans, or safety recalls. The lack of real-world AV deployments means these systems remain largely unproven in real-world threat environments.”

AV safety advocate Dan O’Dowd, founder of The Dawn Project, commented: “Waymo, Baidu and Wayve are bringing autonomous robotaxis to UK roads while Tesla looks in forlornly from outside the party.

“Despite claiming to be on the cusp of solving self-driving for a decade, Tesla’s ‘robotaxi’ is orders of magnitude behind its competitors. Tesla has never given a single driverless public robotaxi ride.”

“Elon Musk is trying to compete in the Tour De France on a tricycle.”