Pony.ai says it will run more than 3,000 robotaxis across over 20 cities in 2026 — more than double a fleet that stood at 1,446 on 25 March, the Guangzhou-based group said. The company reported robotaxi revenue rose 160% year-on-year in the fourth quarter of 2025 and said fare-charging revenue surged more than 500% as total users approached one million. Founder and CEO Dr James Peng, a former Baidu and Google engineer, has framed 2025 as a validation year and is pushing a dual-engine strategy of scaling in China while expanding overseas, amid partnerships with firms including Toyota and Uber; the business remains loss-making despite a recent Nasdaq listing and a valuation above $7bn.

Rapid fleet growth and clear numerical targets

Pony.ai has moved from pilot fleets to substantial commercial scale in under a year. The company said its total fleet reached 1,446 units by 25 March and that it expects to scale to more than 3,000 robotaxis in 2026, deploying across more than 20 cities with nearly half of those overseas.

That growth is driven by a compressed timeline. In public remarks last year, Dr James Peng described the period as an inflection point for the industry and set ambitious targets — more than 1,000 robotaxis on roads from Shenzhen to Dubai soon and a longer-term aim of at least 10,000 vehicles within three years. The company’s March filing presents a more immediate operational target for 2026 while reaffirming the longer-term scale ambitions articulated by Peng.

So far the numbers show momentum. On 22 March Pony.ai’s Gen-7 robotaxi fleet in Shenzhen posted a daily revenue high of RMB 394 per vehicle and averaged 25 orders per vehicle on that day. By mid-February the company said paid Robotaxi orders in Shenzhen for the year had already exceeded the city’s whole-year 2025 total.

Revenue lift and unit-economics milestones

Pony.ai cited a sharp improvement in top-line metrics as evidence that the business model is working at scale. Robotaxi revenue rose 160% in Q4 2025 compared with a year earlier.

Mastheads

Fare-charging revenue climbed more than 500% over the same period, driven by fleet expansion and rising user adoption.

Operationally, the company reported reaching unit-economics (UE) breakeven in Guangzhou in late 2025 and in Shenzhen by February 2026. Those are concrete thresholds for any transport service moving from pilot losses to a repeatable commercial service: breaking even at the vehicle level shows fares and utilisation can cover direct operating costs in certain urban settings.

If Pony.ai can sustain vehicle-level unit-economics while scaling, it would be a clear demonstration that autonomous ride-hailing can move from pilots to repeatable commercial service — a threshold the industry has long sought.

"2025 marked an amazing year for Pony.ai," Dr James Peng said in the company release announcing the results, adding that the firm aimed to replicate the model overseas and build "dual growth engines" in China and abroad. The company also said total users were approaching one million by late March 2026 — nearly three times the user base it reported a year earlier.

Technology, partnerships and a vehicle-agnostic approach

Pony.ai markets itself on software and hardware that can work across different vehicle platforms. That vehicle-agnostic stance helped the firm persuade established mobility and automotive names to test its technology. Dr Peng has highlighted trials and collaborations with Toyota and Uber as part of a strategy to prove Pony’s stack on cars made by other manufacturers.

The company argues that mobility is a very large industry with room for multiple players, and that fast execution and cost-efficient technology can let Pony.ai compete against better-funded rivals. Peng’s background at Baidu and Google is a recurring point in profiles of the company: he frames Pony’s technical depth and adaptability as an advantage in moving from validation to mass deployment.

Global push and the dual-engine strategy

Pony.ai describes its next phase as a dual-engine expansion: deepen operations and unit-economics in China while exporting the model to overseas cities. The March announcement put nearly half of the planned 2026 city rollouts outside China. The company has specifically flagged markets from Shenzhen to Dubai in public comments.

Related Articles

"2025 marked an amazing year for Pony.ai," said Dr James Peng, Founder and CEO of Pony.ai, adding that the company planned to "replicate this model in overseas markets and build dual growth engines to support our next phase of accelerated growth." Pony.ai says it will scale to more than 3,000 robotaxis across over 20 cities in 2026.

This article was created with AI assistance.