When Elaine Zhang hailed an Apollo Go taxi in Wuhan last month, she was pleasantly surprised, finding the Baidu-developed autonomous ride-hailing service “much smarter” than when she first tried it in Beijing back in 2021.
Unlike her earlier experience, this time there was no human driver in the car ready to take control during rush hour.
Zhang, a tech worker in her 30s, was surprised when the vehicle honked as another one tried to pass on the right, and even more so when it accelerated in response to a lane change attempt by another driver. She was also pleased at being able to select her favorite songs to play from the back seat.
“The technology is definitely improving, though I think it still can’t match the intelligence of human drivers. I won’t take it if I am in a rush, as my 3.5-kilometer ride took almost 20 minutes,” she said. “The car moved very cautiously around pedestrians and drove in a pretty stiff way.”
Zhang’s ride is a sign of how far Chinese self-driving players have come in just a few years, but the race to lead this potentially lucrative industry is far from over and the other major global contender, the US, is also upping its game.
In China, Baidu, Pony.ai, and WeRide have gained approvals to operate in tier one cities such as Beijing and Guangzhou, while in the US, Waymo is offering paid robotaxi rides to the public in San Francisco and other cities.
Tesla is expected to finally debut its own robotaxi in Austin, Texas, on June 22, a move that could be a milestone for the company and the industry despite coming later than expected and offering less than initially promised.
In October last year, the US electric vehicle giant unveiled a two-door, silver robotaxi prototype dubbed the Cybercab at a launch event and Musk said at the time that Tesla’s robotaxi would not be equipped with a steering wheel or any pedals. The first batch of robotaxis to be launched in Austin, however, will not be the futuristic Cybercab but rather regular Model Y cars.
RELATED ARTICLE
“The use of Model Y shows Tesla is so desperate to launch the robotaxi service and they are desperate to gather more data for commercialization,” said one executive from a leading Chinese robotaxi company, who asked not to be named.
Musk has said the Cybercab will hit the market “before 2027.”
In China, meanwhile, several companies already have cars on the road. Baidu operates a fleet of around 1,000 Apollo Go robotaxis, which provided more than 1.4 million rides in the first quarter. Pony.ai has a fleet of over 300 robotaxis and aims to expand it to 1,000 vehicles by the end of this year and 2,000–3,000 by the end of 2026. WeRide’s fleet numbers around 400.
Goldman Sachs expects that there will be around 500,000 robotaxis operating across more than ten cities in China by 2030. But the investment bank said, in a note last month, the question now is how companies will commercialize the rapid pace of autonomous development.
The total potential revenue opportunity in China’s robotaxi sector will reach USD 47 billion by 2035, up sharply from an estimated USD 54 million in 2025, Goldman Sachs estimates, with growth driven by falling hardware and algorithm costs, along with lower operating expenses for fleet operators.
Waymo, meanwhile, started rolling out paid robotaxi rides to the public in San Francisco and other US cities last year. The company said in April that it now provides more than 250,000 fully autonomous fare-charging rides every week. Waymo, owned by Google parent Alphabet, has a commercial fleet size of over 1,500 vehicles and expects it to grow to 3,500 by 2026.
It surpassed Lyft as the second largest ride-hailing service in San Francisco as of April, making up 25% of rides ordered in the city, according to market intelligence firm YipitData.
Waymo’s successful rollout there is viewed as the “ChatGPT” moment for robotaxis, according to multiple industry executives and investors, as it has proved that the service has a valid path to commercialization after years of investment.
One hurdle specific to the Chinese market, meanwhile, is social in nature.
Unlike in the US, robotaxis in China are mostly only allowed to operate in designated non-central urban areas in cities such as Beijing, Shanghai, Guangzhou, Shenzhen, and Wuhan. According to a senior executive from another Chinese robotaxi company, Chinese operators are actively working with regulators to expand their fleets into more areas, but one of the main challenges is regulators’ concerns about potential job losses.
“Taxi drivers are struggling as their businesses are impacted by ride-hailing platforms,” he said. “Now, even ride-hailing drivers themselves are finding it tough to make a living amid the broader economic downturn, making regulators more cautious in allowing more expansion at this moment, as social stability is the key.”
Robotaxi operators on both sides of the Pacific are looking abroad for opportunities.
In April, Waymo deployed a fleet of robotaxis in Tokyo that will initially be driven by drivers from Japanese ride-hailing platform Go and Nihon Kotsu, Tokyo’s largest taxi company. This will allow Waymo to collect data and train the vehicles before launching its paid driverless service in Japan.
Baidu, Pony.ai, and WeRide, meanwhile, have all announced plans to expand in the Middle East.
Baidu’s Apollo Go aims to deploy 100 robotaxis in Dubai by the end of this year, with ambitions of increasing that number to at least 1,000 within three years. Apollo Go is also partnering with UAE-based Autogo to deploy its driverless fleet in Abu Dhabi. Both Pony.ai and WeRide are partnering with Uber to deploy robotaxis in Europe and the Middle East.
Tesla, despite being a step behind its Chinese and US peers, may have some advantages of its own, according to Goldman Sachs’ analysts, who said in a note that the company’s existing EV business could help reduce the cost of its robotaxis, a critical consideration as robotaxi operators scale up their fleets.
Dan Ives, managing director at Wedbush Securities, agreed that Tesla’s robotaxi vision could be complementary to its carmaking business, as both depend on artificial intelligence training of the vehicles.
“Waymo will also be a winner and Uber will benefit in autonomous [driving] … but we see the true autonomous winner as Tesla,” Ives said. “We estimate the AI and autonomous opportunity is worth at least USD 1 trillion alone for Tesla.”
And despite the recent blow-up between US President Donald Trump and Elon Musk, Ives said approval of Tesla’s robotaxis and FSD, or full self-driving, autonomous driving system will get “fast-tracked” by the administration, given Trump’s “focus on AI dominance over China.”
Musk could also have a unique opportunity in China, where his company helped lay the groundwork for the EV industry. Last year, Chinese state media reported that Beijing told Musk it would welcome Tesla to do some robotaxi tests in the country, as it could “set a good example.”
If Musk were to succeed in rolling out robotaxis in China, local players would see it as a positive rather than a threat, the above senior executive from the Chinese operator told Nikkei Asia.
“This market is nowhere near cutthroat competition yet. It’s still a blue ocean,” the senior executive said. “Right now, all the players combined haven’t even captured 0.1% of its potential, so there’s no need for concern. If Tesla’s robotaxi enters China and helps grow the overall market, while also advancing policy and increasing public awareness, it would be a huge benefit for the entire industry.”
But not everyone believes expertise in making EVs guarantees success in running a robotaxi service.
Tesla or other carmakers may have their advantages as manufacturers, but they don’t necessarily have the operational knowledge to optimize the robotaxi service, said Jun Peng, CEO of Pony.ai.
While vehicle production cost will be a factor in success, the ability to operate a fleet efficiently, including providing maintenance and customer service, will also be key, especially as companies ramp up service in more locations.
“And if we are talking about cost, that’s Chinese companies’ strengths,” said Peng, pointing to the supply chain ecosystem in China that allows Pony.ai to produce robotaxis at a fraction of Waymo’s cost.
Chinese robotaxi players have more cost advantages in vehicles and autonomous driving kits, thanks to the country’s robust automotive supply chain, including for components like batteries and LiDAR, or light detection and ranging sensors. In contrast, US robotaxi firms have greater access to advanced hardware, such as high-performance chips, and typically enjoy stronger capital support due to backing from major technology and automotive corporations, said Ming Lee, head of Greater China automotives and industrials research at BofA Global Research.
As China and the US each push ahead with their self-driving ambitions, politics will likely play an increasingly important role in the race.
The administration of previous US President Joe Biden issued a ban on Chinese autonomous vehicle software and hardware last year that is set to go into effect in 2026. Meanwhile, Beijing’s data export controls have hindered attempts by foreign companies, including Tesla, to enter the China robotaxi market, though draft guidelines were issued earlier this month to regulate the export of automotive data generated within the country. The proposed rules could offer a pathway for Tesla and others that have been trying to transfer such data abroad to make inroads in China.
In the near future, Europe and the Middle East are expected to be the key markets where Chinese and US players compete head-to-head.
“Low cost is the major advantage for Chinese robotaxi companies,” BofA’s Lee said. “In addition to that, Chinese companies having an edge in advancing their algorithms faster, as the road conditions and drivers’ behavior in China are more complicated, allowing these Chinese robotaxi companies to collect more data in complicated driving scenarios. As a result, these Chinese companies are ready to roll out their robotaxis in overseas markets sooner.”
In China, meanwhile, a sense of urgency is growing. The senior executive from the Chinese robotaxi company said industry leaders have repeatedly urged their government to recognize the progress made in the US, warning that China risks falling behind if companies remain restricted to small-scale trials in peripheral urban areas.
“The Chinese government will still be keenly attentive to US developments in this sector, as only China and the US currently possess the requisite technological capabilities,” he said. “At the policy level, Beijing maintains a supportive stance overall.”
This article first appeared on Nikkei Asia. It has been republished here as part of 36Kr’s ongoing partnership with Nikkei.