11/18 2026
554
Recently, Pony.ai and Toyota's seventh-generation bZ4X was deployed in demonstration operations in Guangzhou, making "full coverage across Beijing, Shanghai, Guangzhou, and Shenzhen" a claim that can now be delivered on. But what truly defines Pony.ai's standing is not the number of cities covered, but a set of account books that are rapidly diverging. What impact will Pony.ai and many other autonomous driving mobility companies have on ride-hailing and taxi drivers?
At the table in first-tier cities, Pony.ai sits in the "profitable seat"
L4 mobility services in Beijing, Shanghai, Guangzhou, and Shenzhen have formed a clear echelon: Apollo Go ranks first in scale with 250,000 weekly orders and over 20 million cumulative trips; WeRide has a global fleet of over 2,100 vehicles, pursuing a "breadth-first" overseas expansion route; Cao Cao Mobility and Ruqi Mobility, backed by Geely and GAC respectively, have entered as automaker-affiliated players.
What makes Pony.ai special is that it is the first player to simultaneously achieve positive per-vehicle operating profit in two first-tier cities, Guangzhou and Shenzhen—Guangzhou averages 23 orders per day at 299 yuan, while Shenzhen averages 25 orders per day with net revenue up to 394 yuan. During the window of transition from a "technology story" to a "business story," this position carries more weight than fleet size.
The cost scissors gap: the ride-hailing driver's rival is not the "driverless car" but the "car that doesn't need to split revenue"
Traditional ride-hailing drivers typically earn daily revenue in the range of 200 to 300 yuan, and after deducting commissions and vehicle costs, net income mostly falls between 3,000 and 5,000 yuan per month.
Pony.ai's Shenzhen fleet has reached daily net revenue per vehicle of 338 to 394 yuan, more than 80 yuan higher than local human-driven ride-hailing, amounting to a revenue gap of nearly 30,000 yuan per year.
More critically, the cost decline curve: the seventh-generation autonomous driving kit costs 70% less than the previous generation, and the 2027 vehicle model targets a total cost within 230,000 yuan; in safety-officer-free mode, the per-kilometer cost can reach 1.13 yuan, compared to approximately 1.4 yuan for electric ride-hailing vehicles. Robotaxi doesn't need to split revenue with the person behind the steering wheel, and this structural advantage will not be reversed by minor fluctuations in technology.
The essence of the employment impact: not "overnight unemployment" but "orders being quietly siphoned away"
Nationwide, licensed ride-hailing drivers already exceed 7.48 million, equivalent to approximately 4.4 to 5.4 million full-time drivers, with a large number of practitioners already in a state of "transitional employment."
Robotaxi substitution therefore exhibits a layered pattern: in some autonomous driving pilot areas, express ride-hailing drivers' daily orders have declined by 15% to 30%, with some drivers in Guangzhou reporting order volumes dropping from 20 to 8–15.
This is not a one-time elimination of jobs, but a continuous squeezing process—each additional Robotaxi order thins the marginal income of human drivers by another layer, until some voluntarily exit. New positions such as remote monitoring and ground operations are growing, but the skills gap between "priority hiring" and "smooth transition" cannot be bridged overnight.
Monopoly and price increases: oligarchs will emerge, but "arbitrary price hikes" won't come easily
The industry's CR5 has reached 68%, and the outline of oligopolistic competition has already emerged. However, the probability of moving toward a single-giant monopoly is not high: mobility is a heavily regulated public service, and pricing power cannot be entirely handed over to capital; technological routes have not yet converged, with automaker-affiliated, tech-affiliated, and platform-affiliated players each holding their own chips, and a landscape of "multiple parties across the industrial chain sharing capital" fits policy preferences better than winner-takes-all.
As for price increases, the more likely path is: low prices to gain scale—scale to build barriers—moderate price increases after barriers are formed: first use prices lower than ride-hailing to weld user habits in place, and after human driver supply contracts significantly and alternatives become scarce, the price center slowly shifts upward.
Whether prices rise by then will not depend on the conscience of capital, but on whether regulators treat Robotaxi as infrastructure to be priced like water and electricity.