Pony.ai released its third-quarter financial report for 2026 on November 18, revealing total revenue of 299 million yuan — a 75.2% year-on-year surge that marks a quarterly record. Notably, Robotaxi revenue jumped 165.4%, with its share of total revenue approaching 40% for the first time. As revenue climbed, operating expenses rose by a modest 3.1%, allowing scale effects to drive efficiency gains and narrow losses.
From an operating zone spanning over 1,000 square kilometers in Guangzhou to Europe' s first fully unmanned Robotaxi passenger test, Pony.ai' s heavy investment in technology is translating into tangible commercial progress.
Robotaxi Share Nears 40%, Operating Leverage Takes Effect
In the third quarter, Pony.ai' s total revenue reached 299 million yuan — a 75.2% annual increase and a 23.1% rise from the previous quarter, setting a new high. This growth was fueled by both Robotaxi and Robotruck businesses, with the former leading the charge.
During the quarter, Robotaxi revenue hit 119 million yuan, soaring 165.4% year-on-year and climbing 47.1% quarter-over-quarter. That sequential growth accelerated further compared to the second quarter.
A shift in revenue structure is drawing attention: the business' s share climbed to 39.9%, nearing the 40% threshold for the first time, up from 26.3% a year earlier. Robotaxi has solidified its status as a core revenue pillar. Passenger fare revenue alone surged nearly 400%, signaling sustained demand for ride-hailing services.

Source: Pony.ai (same below)
This growth stems from two main sources. First, the entire lineup of seventh-generation Robotaxis — developed with Toyota, BAIC, and GAC — has been deployed in fully unmanned commercial operations, laying the foundation for large-scale rollouts. Second, the "co-built fleet" model, both domestic and international, has boosted revenue from vehicle deliveries.
Even as revenue surged, expense growth lagged significantly. Operating expenses rose just 3.1% year-on-year in the third quarter, far trailing the 75.2% revenue jump. Gross profit improved 69.5% to 53.1 million yuan, while operating losses narrowed. On a Non-GAAP basis, operating losses fell by 7.6%.
CFO Wang Haojun noted that this reflects the emergence of operating leverage as Robotaxi operations scale: R&D and operational costs are being spread across a growing fleet and order book.
Micro-level data underscores this shift. Insurance premiums for the seventh-generation models dropped 19% year-on-year, highlighting cost improvements. The company previously disclosed that unit economics in Guangzhou and Shenzhen had turned positive. Meanwhile, the global Robotaxi fleet grew from 2,000 vehicles at the end of the second quarter to 2,768 by September 30, suggesting the unit economic model is poised for validation on a larger scale.
The Robotruck business is also in a phase of continuous expansion. Third-quarter revenue reached 120 million yuan, up 76.3% annually and 34.7% sequentially, driven by growth in freight services and a partnership with Sinotrans. The T9 heavy truck, co-developed with GAC Lingcheng, debuted in Hanover, Germany, in September and rolled off the production line in Conghua by late October.
Pony.ai has deployed autonomous heavy trucks at Shenzhen' s Mawan Port and Jiangmen Port, aiming for regular commercial operations. Its first autonomous light truck, released this year, has entered production and is undergoing testing with logistics clients in China' s Tier 1 cities. Smart solutions revenue came in at 59.4 million yuan, a 3.4% annual increase, with clients spanning low-speed delivery, smart cleaning, and logistics — sectors where demand remains steady.
Domestic Expansion, Overseas Launches: Dual-Track Push for Annual Targets
Operational expansion is advancing along two distinct tracks: domestic and international.
In the domestic market, Pony.ai is prioritizing high-value cities. Its operational area in Guangzhou has exceeded 1,000 square kilometers, connecting Nansha with the city center to enable citywide, cross-district travel. This cross-district and cross-river operation not only drives high-value long-distance orders but also validates the technology' s ability to generalize across different environments.
In Shenzhen, Pony.ai is currently the only company extending fully unmanned ride-hailing services to Bao' an International Airport — a first for a Chinese Tier 1 city. Deployment testing has also begun in Hangzhou, Chongqing, and Zhuhai' s Hengqin district.

This aligns with the company' s assessment of the domestic market during its second-quarter earnings call.
At the time, Founder and CEO James Peng noted that China' s taxi and ride-hailing market exceeds 10 million vehicles, with demand concentrated in Tier 1 and top Tier 2 cities where market size is large and regulation is relatively mature. Increased deployment shortens wait times and boosts retention rates, driving up daily revenue per vehicle and creating a virtuous cycle of scale growth, cost dilution, and improved profitability. Vehicles can share parking spaces and charge or park autonomously, allowing standardized processes to lower the staff-to-vehicle ratio — meaning entering new cities doesn't require a proportional surge in headcount.
Internationally, the "co-built fleet" model is the primary expansion strategy. As outlined in the second-quarter call, Pony.ai supplies vehicles equipped with its virtual driver, while mobility platforms match demand and local partners handle fleet maintenance. The company avoids holding heavy assets overseas; instead, it collects technology licensing fees or revenue shares over the vehicle' s lifecycle, generating high-margin recurring income.
Pony.ai has partnered with Uber on a plan to deploy more than 2,000 Robotaxis across five European cities, targeting high-fare markets in locations like Zagreb, London, and Manchester. Zagreb has already achieved two European firsts: launching the continent' s first commercial Robotaxi service in April and initiating its first fully unmanned passenger test in September. Local users can hail rides via the Uber and Verne apps. By the end of October, the fleet had logged over 200,000 kilometers and completed thousands of orders. To date, Uber and other partners have committed to deploying a total of over 4,000 vehicles.
Founder and CTO Tiancheng Lou introduced the "World Model 2.0" during the second-quarter earnings call. The system uses AI to automatically identify issues, generate solutions, and verify them before deployment, shortening the adaptation cycle for new cities. The same system can generalize across platforms between Robotaxis and Robotrucks, boosting technical reuse efficiency. As of September, the company' s cumulative global autonomous road testing mileage exceeded 120 million kilometers, with fully unmanned mileage surpassing 58 million kilometers.
Regarding annual targets, Pony.ai plans to deploy 3,500 Robotaxis across 20 cities globally by year-end, aiming to lift full-year Robotaxi revenue to 3.5 times 2025 levels. Management previously expressed confidence in exceeding annual revenue goals, a sentiment bolstered by the third quarter' s 165.4% revenue surge and a fleet size of 2,768 vehicles.
Looking ahead to next year, James Peng stated that recent progress in both Robotaxi and Robotruck businesses has strengthened the company' s confidence in driving a significant scale-up of autonomous operations across diverse markets.









