If you walk around the streets of downtown San Francisco and try to hail a taxi, you might notice something unusual: not all of them have drivers. Autonomous cars have been tested on roads in the Californian city for more than a decade, but in recent years, robotaxis have become an increasingly common sight, with […]
If you walk around the streets of downtown San Francisco and try to hail a taxi, you might notice something unusual: not all of them have drivers.
Autonomous cars have been tested on roads in the Californian city for more than a decade, but in recent years, robotaxis have become an increasingly common sight, with initiatives becoming a focal point for mobility innovation.
Alphabet subsidiary Waymo this month completed its 10 millionth journey and now operates in Phoenix, San Francisco, Los Angeles, Austin and Miami, with plans to expand to Washington DC in the coming year.
Meanwhile, the likes of Uber, Lyft and even Elon Musk’s Tesla are planning to deploy their own robotaxi solutions.
But why are automated taxis thriving in some markets, such as the US and China, while others continue to travel in the slow lane?
US accelerating robotaxi plans
Tesla is poised to officially enter the robotaxi race, with plans to launch its autonomous ride-hailing service in Austin, Texas, on 12 June 2025. The pilot will start with 10 modified Model Y vehicles operating without drivers, each supported remotely by teleoperators, initially serving only invited users.
Musk suggests this is just the beginning, with a broader rollout targeting 1,000 robotaxis in the coming months. The company has also teased the unveiling of a purpose-built autonomous vehicle, dubbed the “Cybercab”, set for mass production in 2026.
Yet even Tesla’s own head of autonomy, Ashok Elluswamy, recently admitted that the company’s self-driving technology still lags one of the market leaders: Waymo.
“Technically, Waymo is already performing. We are lagging by a couple of years,” Elluswamy said on the Gobinath podcast.
Waymo, an Alphabet subsidiary, completed its 10 millionth journey this month and currently operates in five US cities, with plans to expand further. Waymo co-CEO Tekedra Mawakana told CNBC the company is providing 250,000 paid trips a week—double the figure from five months ago.
“These are all paid trips, and they represent people who are really integrating Waymo Driver into their everyday lives,” Mawakana said at the Google I/O developer conference.
Bumps in the Road
Adoption in the US has not been without setbacks. General Motors-owned Cruise accelerated testing in 14 cities, including Seattle, San Diego, Miami, Nashville, Raleigh, Charlotte, Atlanta and Washington DC, with Dallas and Houston closest to commercial service.
But Cruise faced a significant setback after one of its vehicles was involved in a pedestrian incident in October 2023. This led to California suspending its autonomous driving permits and a nationwide pause in operations. Cruise subsequently issued a recall for 950 vehicles and laid off a substantial portion of its workforce.
In August 2024, the company recalled its entire 1,194-strong fleet over unexpected braking issues, following a nearly two-year investigation by the National Highway Traffic Safety Administration (NHTSA). In December 2024, GM stopped funding Cruise’s standalone business, folding its work on autonomous vehicles into broader advanced driver assistance systems for personal cars.
Despite setbacks, robotaxis operate in 17 US states, though only seven allow them to run without a backup driver. In China, at least 19 cities are hosting robotaxi services or trials. But in the UK? None so far.
The UK’s stuttering progress
The Automated Vehicles Act 2024 paves the way for fully autonomous vehicles, including robotaxis, to be deployed on UK roads. The previous Conservative government forecast that fully autonomous cars would be “on roads by 2026”, but the Labour government now expects a rollout in the second half of 2027.
While limited self-driving technology is already permitted on UK roads, a human driver must be at the wheel and responsible for the vehicle, even if automated technology is engaged.
Zenzic, created by the UK government and industry to champion the UK Connected and Automated Mobility (CAM) ecosystem, is working to accelerate the self-driving revolution. Francis McKinney, head of technology at Zenzic, says the UK has long believed “the future of transport is automated.” He notes that CAM technologies “have the potential to transform the way our people and goods move,” predicting a hybrid future mixing human-driven and automated vehicles.
Francis McKinney, head of technology, Zenzic
McKinney adds that robotaxis are “certainly a significant part of the future” thanks to their potential to enhance safety, reduce congestion and boost the economy. “The UK’s strategic approach aims to manage this transition effectively,” he says, noting the biggest opportunities lie where “existing provision is low and costs are uncompetitive,” with regulation guiding services where the business case is clear.
Safety concerns
One challenge facing wider adoption is a lack of public appetite for automated taxis. LEK Consulting principal Will Chamberlain points to concerns over safety and trust in the technology. Chamberlain co-authored LEK’s latest Global Mobility Survey, which found that 45% of UK respondents said they would never consider using a fully autonomous taxi.
“Key concerns were around safety and trust in the technology,” he says. “While we saw similar reluctance in the US, places like California and Arizona—where over 250,000 driverless rides are now taken each week—had fewer people saying they would ‘never’ use one. It suggests that seeing the technology in action can help reduce resistance.”
A YouGov 2024 poll found that 37% of Brits would feel “very unsafe” travelling in a driverless car.
Why driverless cars have become public enemy number one
Yet automated taxis are expected to reach British streets eventually, with Uber recently telling the BBC it is “ready to launch” in the UK as soon as the correct regulation is in place. Uber has partnered with 18 automated car technology companies, including Wayve, to trial its service. The company already has automated vehicles on roads in the US, China, Singapore and the UAE.
Confluent, a company supplying real-time data infrastructure, is also working with Uber. “The future of transport is undoubtedly automated, but there’s one key technical challenge that isn’t talked about enough,” says Confluent’s Peter Pugh-Jones.
“Before robotaxis become part of everyday life, we need to solve one of the hardest and most overlooked problems in modern technology: real-time data.”
Pugh-Jones explains that autonomous vehicles must not only “see” the road but process constant data streams from cameras, sensors and maps in real time, making split-second decisions. “It’s one thing to demonstrate this in a controlled test zone,” he says, “but it’s another to make it work reliably across an entire city, especially with unpredictable conditions like roadworks, weather or network outages.”
“While the technology behind self-driving is advancing fast,” he adds, “it’s the data infrastructure that will ultimately determine whether autonomous transport can scale safely. Without reliable, low-latency data streaming to power everything from route optimisation to environmental awareness, robotaxis can’t reach the safety standard required for a mass rollout.”
Taking to the streets
There are also physical challenges, particularly in older cities like London, that are less prevalent in places where robotaxis already operate.
Patrick Jeanbart, head of connected cars at Orange Business, notes that UK capital streets are narrower and busier, with not only other cars—including black cabs—but double-decker buses, cyclists and delivery vans in far greater density than in many American cities.
“It’s a constant stream of unexpected decisions,” he says. “A nightmare scenario for autonomous systems, which need clear patterns and stable environments.” Jeanbart acknowledges that the technology itself is “very good” and improving rapidly, but warns that infrastructure could also hold cities like London back.
“These vehicles need strong, low-latency connectivity and constant communication with cloud systems, traffic infrastructure and other vehicles. They’re not just reacting to what’s in front of them. Without solid 5G and local data processing, they just can’t make decisions fast enough to be considered safe.
“To make this work, 5G also has to be supported by complementary technologies like satellite, ensuring seamless coverage and communication across the network—which could increase the cost of deployment.”
Mark Coates, formerly of Highways England and now VP of the Infrastructure Policy Advancement thinktank at Bentley Systems, agrees that UK roads are “far from ready” to transition to autonomous vehicles, highlighting the need for “functional infrastructure”.
Mark Coates, VP of the Infrastructure Policy Advancement, Bentley Systems
“This entails distinct lane markings, uniform signage, precise digital maps and a comprehensive national strategy for enhancement,” he says. “They also require significant digital capabilities.
Without extensive 5G and edge computing, autonomous vehicles will face challenges responding promptly. To truly commit to adoption, it’s essential to enhance connectivity throughout the entire country, extending beyond just urban centres.”
McKinney at Zenzic says technology within automated vehicles is being developed to overcome connectivity issues: “Autonomy can exist in areas with low connectivity and will need to do so to improve the transport networks of rural areas. There are exciting applications available, especially as connectivity improvements enhance the service.”
Looking ahead
The long-term vision for robotaxis remains transformative. According to ARK Invest, widespread deployment could bring transportation costs down to as low as $0.25 per mile. Counterpoint Research forecasts that the global robotaxi fleet could reach 1.6 million units by 2035.
Pricing remains a question, however. LEK’s Chamberlain explains: “In theory, removing the driver—which accounts for 30–50% of operating costs—and running vehicles almost 24/7 could make fares cheaper. But we don’t expect robotaxis to offer ultra-low-cost travel across London due to congestion. We anticipate pricing will remain close to current ride-hailing fares, as we’ve seen in other markets.”
Another question is whether robotaxis on the roads will even matter if flying taxis take off. While it may sound futuristic, Los Angeles has announced a partnership with Electric Vertical Take-Off and Landing aircraft (eVTOL) provider Archer Aviation for the 2028 Olympic Games.
Flying taxis are “actually a lot closer to reality than many people might think,” says Jeff Hoyle, EVP of Aero, Space and Defence at Expleo. “The UK has even set out its Future of Flight Action Plan, which is a roadmap to integrate eVTOL aircraft into national transport systems by 2030.”
“While flying taxis won’t replace ground transport,” he adds, “they’ll likely help relieve congestion and improve access. If industry and regulators can move together at the right speed, the skies could become part of everyday travel sooner than expected.”
