On March 17, Caocao Mobility set up more than 3,600 virtual pickup and drop-off points for Robotaxis in Hangzhou.
Hundreds of autonomous vehicles are already taking orders across the network. Several institutions estimate that by 2030, at least 3 million autonomous vehicles will be on the road.
Beyond Caocao Mobility, platforms including Pony.ai and Apollo Go are racing to stake out territory.
Autonomous vehicles are slowly but relentlessly eating into this market. On the other side are 29.24 million ride-hailing drivers, wondering where they go from here.
Collective Entry
Caocao Mobility's move did not come out of nowhere.
In December 2025, the company unveiled its "10 years, 100 cities, 100 billion yuan" strategy: five operating centers, 100 cities and 100 billion yuan in transaction volume over the next decade.
Its autonomous vehicles had already been running in Suzhou and Hangzhou since February 2025. The 3,600 virtual stops simply made hailing one a little easier.
Zoom out, and the entire intelligent driving sector has already been accelerating.
In China, by the end of 2025, more than 1,000 Apollo Go vehicles were operating on the streets of Wuhan, with no human onboard.
In February 2026, Pony.ai achieved monthly per-vehicle profitability in Shenzhen, with net daily earnings of 338 yuan. Autonomous vehicles had begun making money.
Across the Pacific, Waymo raised $16 billion in February 2026, reaching a valuation of $126 billion. In San Francisco, it captured 27% of the market, putting intense pressure on longtime ride-hailing giant Lyft.
Jiang Hua, CEO of Ruqi Mobility, said in July 2025: "Any company still hesitating over whether to bet on Robotaxis will soon fall behind, perhaps in just a few years." Everyone is rushing ahead. The problem is that 29.24 million ride-hailing drivers are still standing on the track.
Favorites and Castoffs
In fact, even before autonomous vehicles arrived, those 29.24 million drivers were already trapped in a brutal race to the bottom.
Lao Zhou, 42, was previously a factory supervisor in Dongguan. After losing his job, he drove for ride-hailing platforms for three years to support his family.
Lao Zhou told us that after fuel, rental fees and insurance, he took home less than 4,000 yuan a month. "At first, I could make more than 400 yuan by driving 14 hours a day. Now, for the same hours, I get just over 200." Lao Zhou is one of millions of ride-hailing drivers.
QuestMobile data from July 2025 showed that about 29.243 million people were making a living on ride-hailing apps. Some of those accounts may overlap, but the figure is still enough to show the sector is saturated.
A September 2025 survey released by the China New Employment Forms Research Center was even more painful: ride-hailing drivers were nearly 40 years old on average, 70% had taken the wheel after losing jobs, and 60% were the sole source of income for their families.
They had no choice. "Ride-hailing was all that was left."
But the money they can earn is also very limited. Average daily income is 210 yuan in Zhengzhou and just 182 yuan in Yingtan. In Zhuhai, drivers must work more than 12 hours a day to earn 300 yuan, and take home only 4,000 yuan a month after costs. In Kaifeng, more than 65% of drivers make less than 200 yuan a day.
Chengdu is even more extreme. Of its 220,000 licensed ride-hailing vehicles, only about 80,000 are operating on an average day, leaving an idle rate of 66%. Two-thirds of the cars cannot get enough orders even when they are on the road.
In September 2025, drivers asked Didi at a forum why their take-home pay had fallen even though the platform's commission rate had dropped. Didi felt wronged, saying its 2024 commission rate was only 14%, far more reasonable than Uber's in the United States.
The platform was not lying, because the truth is not in the commission rate.
The truth is that, as of July 2025, 29.24 million ride-hailing drivers were fighting over an order pool that had shrunk by 23%.
The number of people rose 23%, while orders fell 23%. That is why almost every ride-hailing driver will tell you they are more exhausted and making less money. They already had no way out, and the arrival of autonomous vehicles has blocked the last path to survival.
Lao Zhou has seen videos of Apollo Go. After subsidies, a 10-kilometer ride costs only 4 to 5 yuan. He charges 18 yuan or more for the same distance. "They don't need sleep, don't need fuel, don't get emotional and are four times cheaper than me. Who do you think the platform will choose?"
Once users get used to a 4-yuan ride, who will still want to pay 18 yuan?
This "new species" needs no social security contributions, runs 24 hours a day and is far cheaper. It exists on an entirely different plane from ride-hailing. From the moment autonomous vehicles appeared, the favored child of this era has stood under the spotlight.
Ride-hailing drivers, meanwhile, are being pushed into a dim corner of history, like castoffs destined to be abandoned.
The door into traditional ride-hailing is closing.
Sanya, Changsha and Shanghai led the way in 2023. Jixi, Tongren, Gaomi and Zhenjiang followed in 2025. By 2026, Ziyang, Taihu County and Aksu had also shut the door.
By early 2026, more than 19 cities across China had stopped issuing ride-hailing licenses. The reason was strikingly consistent: saturated capacity.
In Harbin, anyone who wants to drive for ride-hailing now has to enter a lottery. In November 2025, 1,500 individual quotas drew more than 10,000 applicants, with a success rate below 14%, making it harder than winning a Beijing license plate.
But another door, the one for autonomous vehicles, has been thrown wide open.
While drivers worry about quotas, autonomous vehicles have already received legal licenses.
On December 15, 2025, the Ministry of Industry and Information Technology approved the first batch of L3 access, covering the Changan Deepal SL03 and BAIC Arcfox Alpha S.
Five days later, Chongqing issued China's first official L3 license plate. Three days after that, Beijing issued its first batch of L3 highway plates.
In March 2025, Pony.ai obtained an operating permit for autonomous vehicles in Shenzhen's Nanshan District. By the end of 2025, its operating area had expanded from 21.7 square kilometers to 167.4 square kilometers.
In Shenzhen, functional autonomous vehicles for cargo transport are approved almost instantly and are not constrained by traditional capacity quotas. By December 2025, 1,218 had been deployed.
At the same time, national infrastructure funding is pouring into Beijing, Wuhan and other cities at the scale of tens or hundreds of billions of yuan. In July 2024, 20 cities were selected for the "vehicle-road-cloud integration" pilot program.
Vehicle-road-cloud means cameras at intersections help cars read traffic lights, roadside radar helps cars see blind spots, and a cloud-based brain helps calculate which routes are not congested. The car does not have to do all the hard work itself; the road tells it what to do.
The money is earmarked: 9.94 billion yuan for Beijing, 12.7 billion yuan for Changchun, 17 billion yuan for Wuhan, and more. Governments are spending heavily to build a "smart road" specifically for autonomous driving.
Where Do They Go From Here?
This stark contrast makes one thing clear to everyone: the wheel of history is rolling forward, and no force can stop it.
A speck of dust from the times, when it falls on an ordinary family, becomes a mountain too heavy to bear.
In early 2026, a survey by the Guangzhou Ride-Hailing Industry Trade Union Federation found that more than 80% of drivers feared being replaced.
CPPCC member Fang Ou said the solution is not to halt technology, but to help drivers transition.
That is true. But can those new jobs really absorb these 29.24 million people?
Some say autonomous vehicles still need people, don't they?
Safety operators, maintenance staff, data annotators and remote monitors are all new jobs, aren't they?
But we need to do the math clearly.
As of March 18, 2026, Pony.ai had reached a ratio in Guangzhou where one back-office worker managed 20 vehicles, and it was moving toward a target of 30 vehicles per worker.
By 2030, according to forecasts from Frost & Sullivan and Soochow Securities, Robotaxi penetration in China will reach 30%, with nearly 3 million autonomous vehicles on the road. That also means 3 million taxis that do not need drivers.
Even if every possible new job is counted, at a vehicle-to-worker ratio of 30:1, 3 million vehicles would need only 100,000 people. Add some related roles, and the total may still be only several hundred thousand to one million.
One hundred thousand new jobs against 29.24 million drivers. That cannot even absorb the surplus. More painful still, veteran drivers are unlikely to get even those jobs.
Today's ride-hailing drivers are nearly 40 years old on average, and more than half have only a high school education. Asking them to put down the steering wheel, earn an "intelligent vehicle operations and maintenance" certification, and compete with graduates in their 20s on computer skills is simply beyond what many can handle.
The pain of the old giving way to the new has already arrived. In Wuhan, a taxi driver deliberately staged two collisions with autonomous vehicles in July and August 2025 and received 3,900 yuan in compensation.
There was no person in the autonomous vehicle, but there was 360-degree data evidence with no blind spots, collected in full. The truth was quickly uncovered, and the driver was instead fined 2,000 yuan.
Across the Pacific, autonomous vehicles in San Francisco have also faced protests and blockades. In February 2024, a Waymo vehicle stopped in the middle of the road was set on fire, as bystanders cheered and applauded.
In January 2026, ride-hailing drivers also took to the streets to protest Waymo's expansion. (Image source: The San Francisco Standard, sfstandard.com) The emotion behind these incidents is shared: anxiety over human survival.
Lao Zhou said he has never opposed technological progress. He is just afraid that, in a future filled with celebration, people like him will be brushed aside and written off as the cost of progress.
The wheels of technology can be cold, but civilization should not be indifferent.
The ultimate meaning of any progress is to let more people live with greater dignity.
And to let every person struggling to live be gently held by the times.
Comments
00No comments yet. Be the first to weigh in.