Two Reconciliation Statements for Autonomous Taxi

Invest in stocks? Just look at the Golden Qilin analyst research reports—authoritative, professional, timely, and comprehensive. Help you uncover high-potential themes and opportunities!

Source: Beijing Business Daily

On the road to exploring commercial models for autonomous driving, the two companies have given entirely different answers.

Recently, Pony AI and WeRide have both released their 2025 annual reports, the first annual reports after the “autonomous driving two giants” completed their dual primary listing structures in both the U.S. and Hong Kong. Pony AI achieved break-even profitability for each Robotaxi (autonomous driving taxi) in Guangzhou and Shenzhen; in Shenzhen, the seventh-generation model delivered a highest daily net revenue per car of 394 yuan. WeRide saw its Robotaxi business revenue grow 209% year over year in 2025, while also laying out multiple scenarios such as small buses, sanitation, and freight. Its vehicles have been deployed in 12 countries and 40 cities.

In response to the results, both companies’ executives focused their key interpretations on the Robotaxi business. On the earnings call, Han Xu, founder, chairman, and CEO of WeRide, discussed the outlook: the company’s global Robotaxi fleet target is 2,600 units by the end of 2026—starting phase of its broader long-term vision to deploy tens of thousands of Robotaxis globally by 2030. At a media communication session, Wang Haojun, CFO of Pony AI, said plainly that the per-car profitability model of Robotaxi is the core, and the commercialization inflection point for autonomous driving has arrived.

WeRide’s Robotaxi is driven by a light-asset model, while Pony AI promotes an in-house plus co-building approach. This 2025 annual report has two routes directly colliding head-on. Which side will reach the finish line first in autonomous driving commercialization? The answer is beginning to emerge, but the race is far from over.

Both set new highs

In 2025, both companies’ performance growth curves simultaneously reached historic peaks.

WeRide’s full-year revenue was 695 million yuan, up 89.6% year over year. In the fourth quarter of 2025, revenue increased 123% year over year to 314 million yuan, setting a new record for the highest single quarter since the company was founded. Pony AI’s revenue in 2025 was 629 million yuan, up 20%.

The surge in Robotaxi business is the biggest highlight of both companies’ earnings reports.

In 2025, Pony AI’s Robotaxi ride-hailing service revenue was 16.607 million USD, up 128.6% from 2024. Within that, passenger fare revenue from Robotaxi ride-hailing services grew by about 400% year over year.

In 2025, WeRide’s Robotaxi business revenue was 150 million yuan, up 209.6%; in the fourth quarter of 2025, revenue grew to 50.6 million yuan, up 66.4% year over year.

Profitability samples were also disclosed in the annual report. In November 2025, Pony AI’s seventh-generation Robotaxi began large-scale market deployments. In November 2025 and March 2026, Pony AI achieved per-car break-even profitability within the entire city in Guangzhou and Shenzhen, respectively.

“On average, in February 2026, Shenzhen’s per-car daily revenue was 338 yuan/car, with an average of 23 orders per car per day; at the March peak, Shenzhen’s highest per-car daily revenue was 394 yuan/car, with a highest number of 25 orders per car per day.” Wang Haojun provided a detailed introduction to Pony AI’s latest operating data to a reporter from Beijing Business Daily.

At the earnings call, Han Xu disclosed WeRide’s details: “In the past six months, the average number of daily orders per car was 15; during peak periods, it rose to 26. The average passenger waiting time fell to within 10 minutes. In the fourth quarter of 2025, Robotaxi registered users grew by more than 900% year over year.”

However, neither company achieved profitability. Excluding investment gains, Pony AI’s net loss under non-GAAP accounting standards in 2025 was 174 million USD, expanding 31.5% compared with 2024. WeRide’s 2025 net loss was 1.655 billion yuan, narrowing 34.2% compared with 2024; its adjusted net loss was 1.247 billion yuan, expanding 55.5% compared with 2024.

Sell vehicles or sell the model

From the perspective of revenue structure, the businesses of the two companies cannot be directly compared.

WeRide’s revenue comes in two parts: products and services. Among the 695 million yuan revenue in 2025, 52.5% came from product revenue—both the company’s largest revenue source and the fastest-growing revenue segment. That figure was up 310.3% from 2024, with growth mainly driven by increased sales revenue from autonomous driving taxis, autonomous driving minibuses, and driverless sanitation vehicles. It can be understood simply as “selling vehicles.”

In 2025, WeRide’s service revenue grew 18.8% from 273 million yuan in 2024 to 325 million yuan, accounting for 47.4% of total revenue. This segment has a more diversified revenue mix. Growth was mainly driven by a 104 million yuan increase in intelligent data service revenue and a 17.7 million yuan increase in autonomous driving related operating and technical support service revenue, partially offset by a decrease of 70.10 million yuan in ADAS (advanced driver-assistance systems) R&D services.

WeRide’s Robotaxi revenue of 150 million yuan in 2025 includes both product revenue and service revenue.

On the earnings call, WeRide’s CFO and head of international business, Li Xuan, stated clearly: “Robotaxi is the core business engine, and growth is driven by a light-asset model. We provide the autonomous driving ‘brain,’ while the mobility platform and asset partners are responsible for operations and vehicle ownership. The vehicles are not included in the company’s balance sheet, enabling efficient expansion and ensuring a consistent user experience. Revenue sources include ongoing service fees and ride-fare sharing.”

Pony AI’s model is different. “It’s similar to Didi’s model—Pony AI provides the service, and passengers pay the fares. Part of the Robotaxi also comes from licensing fees and project-based revenue. We believe that with large-scale deployment of Robotaxi, the most important revenue must be passenger fares and the authorization revenue under the co-building model.” Wang Haojun explained.

Returning to revenue structure, Pony AI classifies its revenue by Robotaxi, Robotruck (autonomous driving truck service), technology licensing, and applications. Robotruck is still the biggest contributor to Pony AI’s revenue, achieving 40.601 million USD in revenue in 2025, accounting for 45.1% of total revenue. The share of Robotaxi revenue increased from 9.7% in 2024 to 18.5% in 2025; the annual increase of 128.6% is the largest among the three business segments.

“Robotaxi is the most important.” Wang Haojun told a reporter from Beijing Business Daily, “Eighty percent of the technology between Robotruck and Robotaxi can be shared. Generally, in our next-generation Robotaxi, the corresponding hardware and technologies will be applied to the next generation of Robotruck.”

After the inflection point: the scale race

The financial report numbers are just the starting point. The envisioned scale-up is only beginning, and both autonomous driving two giants have set clear targets.

Based on Robotaxi’s total revenue in 2025, fleet operating scale, coverage footprint, and growth of users themselves—especially the fact that the seventh-generation vehicles achieved UE (unit economics model) break-even in two cities—Pony AI believes that the inflection point has arrived.

The emphasis on the importance of UE break-even lies in that this metric indicates whether the Robotaxi business model can truly work. Wang Haojun further interpreted: “If we ensure scalable, standardized operation of Robotaxi by guaranteeing safety, performance, and reliability, and we optimize the operating processes, then Robotaxi’s costs can be covered by GMV (gross merchandise value).”

In a growth-cycle chart he shared, 2016–2024 is the period of deep accumulation of Robotaxi technology; 2025 shows an inflection point as a commercialization validation phase; and in 2026, it enters a phase of rapid growth. In 2025, Pony AI’s Robotaxi fleet reached 1,446 vehicles. In 2026, it plans to expand the fleet to over 3,000 vehicles, with deployments of more than 20 cities worldwide. Robotaxi business revenue is expected to grow by more than three times. This goal will be achieved through both in-house fleets and co-built fleets.

Currently, WeRide’s autonomous driving products have been deployed in 12 countries and more than 40 cities. Its Robotaxi fleet size has reached 1,125 vehicles. According to the plan, by the end of 2026 the company will deploy 2,600 Robotaxis globally, and by 2030 it will deploy tens of thousands of Robotaxis.

“We recently renewed our agreement with Geely Remote. In 2026, we will add 2,000 upgraded GXR Robotaxis. By the end of 2025, the fleet has already exceeded 1,000 units. Combined with phased delivery of GXR and the retirement of older vehicles, the global Robotaxi target for end of 2026 is about 2,600 units.” Li Xuan provided a detailed introduction to the plan.

According to the Middle East market plan shared by Han Xu, “At present there are about 200 vehicles in the Middle East. The company and Uber plan to add at least 1,000 vehicles by 2027, with the goal of becoming the first company in that region to reach a scale of a thousand-plus Robotaxis.”

The capital markets have priced the two routes differently. As of the close of Hong Kong stocks on April 1, Pony AI was quoted at 73.2 HKD per share, with a total market cap of 31.631 billion HKD. WeRide was quoted at 20.72 HKD per share, with a total market cap of 21.083 billion HKD.

Beijing Business Daily reporter: Wei Wei

		Sina Statement: This news is republished from Sina’s partner media. Sina.com publishes this article for the purpose of transmitting more information and does not mean it agrees with these views or confirms the descriptions. The article content is for reference only and does not constitute investment advice. Investors act at their own risk.

Massive information, precise interpretation—everything is in the Sina Finance APP

Editor: Gao Jia