Pony AI (NASDAQ:PONY) used its fourth-quarter and full-year 2025 earnings call to highlight rapid growth in its Robotaxi business, progress toward profitability at the unit level in key Chinese markets, and an aggressive expansion plan that management said will extend to more than 20 cities globally by the end of 2026.
Robotaxi growth and unit economics milestones
Chairman and CEO Dr. James Peng said 2025 was defined by “remarkable milestones,” led by accelerated top-line growth and commercial scaling following the April debut of its Gen-7 Robotaxi. Peng said fourth-quarter Robotaxi revenues rose sharply year over year and that fare-charging revenue growth was even stronger, which he tied to broader commercial deployment and higher utilization.
Both Peng and Wang emphasized a key operational milestone: positive unit economics (UE) in Guangzhou and Shenzhen. Wang said the company achieved consecutive positive UE in both cities within four months of the Gen-7 launch, citing cost advantages in vehicles and operations and the ability to deliver a consistent, high-quality 24/7 service in complex urban environments.
Management pointed to improving operating metrics in Shenzhen as fleet density increased. Wang said that in February 2026 the company delivered positive unit economics in Shenzhen with an average of 23 daily orders and RMB 338 in average daily net revenue per vehicle. He added that in March the company hit a new daily peak of RMB 394 net revenue and 25 orders per vehicle, and that paid orders in the first two months of 2026 in Shenzhen had already surpassed the full-year 2025 order volume in that city.
Scaling the fleet and entering new cities
Peng said the company’s Gen-7 Robotaxis moved “straight into mass production and commercial deployment,” and that the fleet has surpassed 1,400 units. He reiterated a target to grow the fleet to more than 3,000 Robotaxis, and said the company now has “greater visibility” to potentially exceed that goal.
On the Q&A, Peng said the company has been focused on production of the Toyota bZ4X-based Gen-7 model, while continuing production with Beijing Auto and Guangzhou Auto. With those three vehicle programs, he said Pony.ai is confident it can reach “over 3,000 units by the year-end,” while expanding deeper into downtown hubs and launching in additional markets.
Peng also detailed the company’s geographic expansion. In China, he cited further penetration in Shenzhen traffic hubs during Chinese New Year, entry into University Town in Guangzhou, and new launches in Hangzhou and Changsha. Internationally, he described progress in Europe, the Middle East, and Southeast Asia, and said the company’s overseas footprint now covers a population of 100 million.
Global partnerships and the joint deployment model
A central theme of the call was the company’s “dual engine” strategy, which management described as pursuing growth in both China and international markets. Peng said the company plans to deploy Robotaxis in more than 20 global cities by the end of 2026, and he told analysts that nearly half of the targeted cities for expansion are expected to be outside China, spanning Asia, Europe, and the Middle East.
Management repeatedly referenced strategic partnerships as enablers of expansion. Peng cited alliances with Tencent and Uber, including an integration with WeChat Mobility intended to make it easier for users to call Pony.ai’s Robotaxis. Overseas, he cited partnerships with Uber and also regional relationships with Bolt and Stellantis, as well as a collaboration with Uber and Verne (a Rimac Group company) to enter Croatia, which he characterized as Europe’s first commercial fare-charging Robotaxi service.
Executives also described a “joint deployment model” designed to improve capital efficiency. Wang said partners fund vehicle capex and participate in the Robotaxi value chain (including ground operations, maintenance, and charging), while Pony.ai expands in a more asset-light way. He said the company expects nearly half of new vehicles added this year to come through this model, led by Toyota, and that Pony.ai can earn recurring revenue through revenue sharing or an AI driver license fee in addition to fare-charging revenue from its self-owned fleet.
When asked about geopolitical tensions in the Middle East, Peng said the region remains a high priority and that the company had not seen any material impact from current tensions. He said the company expects to roll out fare-charging services with Mwasalat in Doha and is preparing for fully driverless operations in Dubai following approval “later this month.”
Technology focus: world model, cost reduction, and operating moat
CTO Dr. Tiancheng Lou framed 2025 as a “landmark year” for proving commercial viability and argued that technology leadership is now tied directly to scale, efficiency, and user experience. He said Pony.ai’s “world model” approach is a widely recognized path for autonomous driving and that the company’s system is designed to generalize across cities because “driving is about interaction and negotiation” with other agents, with differences between cities largely reflecting different distributions of similar scenarios.
Lou said the world model helps accelerate entry into new markets by generating large-scale simulated scenarios that reflect traffic patterns in a new city, enabling reinforcement learning and more efficient validation and fine-tuning without collecting massive amounts of new real-world data. He also said the company’s broad operational design domain (ODD) and standardized operating workflows support deployment beyond limited, low-complexity routes.
On efficiency, Lou said the company is aiming for a “clear cost advantage” and described plans to lower Gen-7 bill-of-materials costs via more cost-effective components, while leveraging safety performance to reduce insurance fees and improve remote assessment efficiency. Wang added that the company had proactively secured procurement for critical components, including memory modules, and said it expects minimal impact from supply chain price fluctuations. He also stated the company remains on track to achieve a 20% reduction in ADK bill-of-materials costs in 2026 versus 2025.
Asked about NVIDIA’s open-source autonomous driving model announcements, Lou said Pony.ai views ecosystem progress as beneficial, but emphasized the gap between a model and a commercially deployed Robotaxi fleet with proven safety, regulatory approvals, and scaled operations. He said Pony.ai’s advantage comes from “full stack in-house development” combined with real-world L4 deployment, including vehicle architecture, sensors, redundancy, domain controllers, validation, and commercialization.
Robotruck and licensing/applications updates, plus profitability and capital
Beyond Robotaxis, Peng and Wang discussed progress in Robotruck and licensing/applications. Peng said the company introduced its Gen-4 Robotruck in 2025 and reduced ADK bill-of-materials cost by 70%, with a goal to mass-produce and deploy Gen-4 Robotrucks. Wang added that an EV transition is expected to lower per-kilometer operating costs and said the company expects accelerated Robotruck revenue growth beginning in the second half of the year as it expands route coverage and scenarios such as dedicated lines and port operations.
On licensing and applications, Peng said the company’s autonomous domain controller (ADC) sales grew substantially, and Wang stated ADC volume rose to six times the level of 2024. Management also cited expansion into low-speed delivery, robot sweepers, logistics, and robotics.
Wang said the company achieved a “first-ever quarterly GAAP level net profit” in the fourth quarter, attributing the result primarily to gains from strategic equity investments. He added that 2025 expenses widened slightly due to front-loaded investment in Gen-7 mass production, new city expansion, and strengthening the tech stack.
Management closed by emphasizing liquidity following its Hong Kong IPO. Wang said the company ended the year with more than $1.5 billion in cash reserves as of Dec. 31, 2025, including IPO proceeds of more than $800 million. He said the company expects to increase investment in R&D, AI talent, go-to-market capabilities, and international expansion, while aiming for revenue growth to outpace operating expense growth as fleet scale increases and positive unit economics create a “virtuous cycle.”
About Pony AI (NASDAQ:PONY)
Pony.ai develops autonomous driving technologies for passenger and goods transportation. The company offers an end-to-end self-driving stack that combines perception, planning and control systems with proprietary hardware and software. Pony.ai’s solutions support robotaxi services and advanced driver-assistance system (ADAS) deployments across urban and suburban environments.
Founded in late 2016 by James Peng and Sean Gong, Pony.ai operates research and development centers in Fremont, California, as well as in Guangzhou and Beijing, China.
