Robotaxi is entering the "China Moment"
```
On December 21, Waymo, the global leader in Robotaxi, attracted attention for large-scale stoppages at multiple intersections.
The cause of this incident was the malfunction of traffic lights in San Francisco due to a power outage.
Waymo’s Robotaxi was unable to flexibly respond to unexpected situations like extinguished traffic lights, opting instead to park at intersections—a seemingly “safer” maneuver.
This incident also serves as a warning for domestic Robotaxi fleets poised for rapid expansion.
Looking back at all of 2025, the capitalization of China’s Robotaxi track has been accelerated.
According to incomplete statistics from Xinfeng, several companies—including Hello, Didi, Pony.ai, and WeRide—have completed funding rounds this year, with total financing exceeding ten billion yuan.
This has provided ample fuel for the expansion of each company's Robotaxi fleet in 2026.
Based on publicly disclosed target fleet sizes for 2026, Xinfeng estimates that Pony.ai, WeRide, and Didi alone may collectively reach a fleet of 8,000 Robotaxis in 2026.
China's Robotaxi is approaching the milestone of 10,000 vehicles.
In contrast with its competitors, Baidu’s Apollo Go, China’s No.1 Robotaxi, has been extremely low-profile.
To date, Apollo Go remains merely a department of Baidu, without independent external financing, leaving much room for speculation.
Behind this “restraint,” Apollo Go is also planning to expand its operational scale. Recently, Robin Li stated during the earnings call a plan to increase vehicle deployment in existing cities and expand to more new cities in 2026.
Although various parties believe that 2026 will be the inaugural year of large-scale Robotaxi deployment, many challenges remain unresolved.
On one hand, scaling up Robotaxi fleets will expose the strengths and weaknesses of each company’s technological stability; whoever can truly control rates of injury incidents and average intervention mileage may hope to win this war;
On the other hand, as the transition from “human-machine co-driving” to “fully driverless” is key, the implementation of L3 autonomous driving also adds urgency to the expansion of the Robotaxi industry.
Rise of the Warlords
The biggest change in China’s Robotaxi sector in 2025 is the increasing number of entrants, with the market entering a competitive landscape akin to ‘warlords rising together’.
Previously, participants in the domestic Robotaxi sector were primarily focused around three autonomous driving tech companies: Apollo Go, Pony.ai, and WeRide.
Now, ride-hailing platforms and car manufacturers with user traffic entry points are joining the fray.
Currently, domestic Robotaxi players fall into three main categories:
First are autonomous driving tech companies, represented by Baidu Apollo Go, Pony.ai, WeRide;
Second are ride-hailing platforms, represented by Didi, Ruqi Mobility, Caocao Mobility, Hello, and AutoNavi;
Third are major manufacturers, represented by XPeng and others.
At present, these three categories of players have fostered a complex relationship of competition and cooperation, manifesting a triangular model of ‘AI autonomous driving company algorithms + mass-produced cars from manufacturers + ride-hailing platform order network’.
For example, Pony.ai has co-developed the seventh-generation Robotaxi model with BAIC Arcfox, GAC Aion, and Toyota; Didi is working with GAC Aion on L4 Robotaxi development;
XPeng’s Robotaxi has opened SDK interfaces to AutoNavi, with the latter expected to handle future operations.
Behind this is the existential anxiety of Chinese Robotaxi companies.
In China, whether it is independent autonomous driving tech firms struggling with losses or manufacturers facing sales pressures, neither can single-handedly sustain the years-long capital-intensive Robotaxi “money-burning game.”
In contrast, Waymo, the global Robotaxi leader, is backed by Google, providing virtually limitless capital—allowing it to shoulder full-chain burdens from algorithm R&D to heavy asset operations.
Even so, Waymo is still planning external financing. Market reports say Waymo is negotiating a new round of funding at a valuation of $100 billion.
Against this backdrop, cooperation is the optimal solution for Chinese players to share risks and allocate resources.
Supply chain-based cost advantages are an important strength of Chinese Robotaxi players.
Pony.ai CFO Wang Haojun believes that cost is the chief advantage of Chinese companies vis-a-vis Waymo.
“If we look at Pony.ai’s seventh-generation against Waymo's Zeekr or Hyundai models to be deployed this year, we see their cars are more than 4–5 times more expensive than ours. So, if competing in overseas markets, Chinese companies will have significant cost advantages,” said Wang.
He further explained to Xinfeng, “From Waymo’s perspective, they don’t need to tell Google that they must quickly reach single-vehicle profitability before expanding. Even if they deploy thousands of vehicles and lose billions, it’s not an issue for Google. Under this mindset, we each approach cost control differently. Also, today China’s supply chain can achieve hardware costs much lower than the U.S. supply chain.”
Referencing Waymo’s growth path, investing as many Robotaxis as possible to achieve scale effects and dilute unit operating costs is at the core of achieving commercial viability.
10,000 Vehicles Imminent?
Major funding rounds in the autonomous driving track this year hint at looming large-scale battles.
In the second half of the year, Hello’s Robotaxi business and Didi Autonomous Driving each secured 3 billion and 2 billion yuan in funding; subsequently, Pony.ai and WeRide listed on the Hong Kong Stock Exchange on the same day, raising 6.1 billion and 2.2 billion yuan respectively, totaling over 13 billion yuan in financing.
Breaking down the purpose of these raised funds, most are earmarked for Robotaxi fleet expansion.
According to Xinfeng’s incomplete statistics, Pony.ai, WeRide, and Didi’s 2026 Robotaxi fleet goals are 3,000, 2,000–3,000, and 1,000–2,000 vehicles respectively, for a total approaching 8,000.
China’s Robotaxi fleets may be very close to the milestone of 10,000 vehicles.
In summary, the reasons for fleet expansion in 2026 can be boiled down to three:
First, per-vehicle cost reduction this year has yielded more profit potential;
The domestic company with best per-vehicle cost performance is Apollo Go.
Apollo Go’s sixth-generation unmanned vehicle now costs 204,600 yuan, 60% less than the fifth generation—about one-seventh of Waymo’s cost and 70% of Pony.ai’s seventh-generation cost; Pony.ai’s seventh-generation autonomous kit BOM cost is 70% lower than its sixth generation.
Declining hardware costs underpin rapid fleet expansion.
Second, some autonomous tech companies have achieved single-vehicle profitability in specific cities, preliminarily verifying scale profit possibilities;
This year, Apollo Go and Pony.ai reached breakeven per-vehicle in Wuhan and Guangzhou respectively, finally escaping the “lose money on every additional car” situation; WeRide is about to achieve the same in Abu Dhabi.
Third, regulatory policies are relaxing.
This December, Nansha District, Guangzhou, officially released the “Road Opening Implementation Guide for Unmanned Equipment (Trial),” implementing graded management for autonomous equipment across the district and became the first administrative region in the city to announce open roads without application requirements.
This means eligible Robotaxis can operate in approved areas with greatly simplified processes, reducing compliance and time costs.
HSBC expects regulators to issue more licenses and open up more service areas. In Guangzhou, it’s anticipated Robotaxi service will soon cover the whole city—not just limited areas like Nansha.
Yet amid this investment boom, Apollo Go remains an exception.
In terms of per-vehicle cost, fleet size, or mileage accumulation, Apollo Go is unequivocally China’s No.1 Robotaxi operator.
Apollo Go currently spans 22 cities worldwide, including Beijing, Shenzhen, Wuhan, Chongqing, Shanghai, Hong Kong, Dubai, and Abu Dhabi.
As of the end of October this year, Apollo Go’s total service orders exceeded 17 million, with cumulative mileage over 240 million kilometers, of which fully driverless mileage surpassed 140 million kilometers, making it the only Chinese Robotaxi company approaching Waymo’s fully driverless mileage.
While Robin Li announced plans to increase vehicle deployment and expand into more cities in 2026, Apollo Go has still not sought external funding for fleet expansion.
Baidu insiders confirmed to Xinfeng that Apollo Go remains simply a Baidu business unit, with no independent external financing.
This may relate to Baidu’s greater internal funding capability. Compared with independent Robotaxi players like Pony.ai and WeRide, Apollo Go backed by Baidu can clearly access more resources.
But whether Apollo Go can be first to achieve commercial viability without much external funding, amid the need for large-scale capital investment, remains to be seen.
Singularity Moment or Another Experimental Year?
Amid the surge in investment and fleet expansion, it is still unclear whether 2026 will be the Robotaxi singularity moment or just another experimental year; the sector remains shrouded in both technological and commercial uncertainty.
Despite recent fundraising and vehicle deployment seemingly laying the groundwork for scaling up, crossing from trial to full urban rollout still faces many uncertainties.
The greatest variable comes from technical stability.
Even with policy support allowing Robotaxi to operate outside pilot zones, entering old urban areas with complex, unstructured road conditions means autonomous systems will face exponentially greater random risks.
For example, the absolute number of accidents may rise as the fleet size scales up.
On December 6th, a Hello Autonomous test vehicle in Zhuzhou, Hunan was involved in an accident, dragging a pedestrian under the car. The business has since been suspended and the company is cooperating with the investigation.
This incident casts doubt on Hello Robotaxi's claim that it can "cover in two years what others did in ten."
Large-scale operations may also create more friction with cities. For example, on December 21st, a Pacific Gas & Electric substation fire caused a widespread blackout, disabling traffic lights in San Francisco. Local Waymo vehicles could not perceive the failed lights and were stranded in the middle of intersections, causing traffic jams and trapping passengers inside.
From a longer-term perspective, the Robotaxi industry may also face competition from passenger vehicles.
Robotaxi must compete not only with traditional ride-hailing platforms like Didi, but also with privately owned cars.
As early as 2016, Elon Musk raised the idea of turning Tesla vehicles into income-generating shared fleets. Ten years on, Tesla is nearing this vision and has started driverless Robotaxi road tests in Austin, Texas, with no human occupants.
Meanwhile, there are signs of progress in China.
Recently, the Ministry of Industry and Information Technology officially announced the first batch of L3 conditional autonomous vehicle permits: BAIC Arcfox and Changan Deep Blue each have one pure EV model entering pilot testing in designated areas of Beijing and Chongqing.
This landing reconstructs the responsibility chain for driving; according to the national “Automated Driving Classification” standard, L3 places driving control and liability on the system itself under certain conditions.
Many industry insiders believe that although it is impossible to predict when L3 to L4 on passenger cars will take off en masse, this L3 rollout has exceeded expectations and gives hope for the transition to L4 in consumer autos.
This makes the vision of “hands off, eyes off driving”—even “my car works as a Didi while I'm at work”—much more imaginable.
However, industry insiders remain cautious over whether these L3 advances in passenger cars will impact Robotaxi.
“Private car scenarios are relatively often point-to-point and suitable for broader settings. In the short term, I don't think it will impact Robotaxi—it should be more complementary and integrated. In the long term, though, competition may emerge,” GuanGai (Shanghai) Technology CEO Liu Bo told Xinfeng.
Risk warning and disclaimerThe market is risky, and investment must be cautious. This article does not constitute individual investment advice, nor does it consider any particular user’s special investment goals, financial circumstances, or needs. Users should consider if the comments, opinions, or conclusions herein fit their specific situation. Investments made based on this article are at one’s own risk. ```