Q & A Highlights

Q: Given that Uber partners with several autonomous driving companies worldwide, what are the main reasons that made Uber choose Pony.ai in its European rollout? A: James Peng (CEO): Uber chose Pony.ai because we offer reliable technology at scale and an attractive cost structure. Our Gen-7 robotaxis have proven their capabilities through large-scale commercial operations in all Tier 1 cities in China, with positive unit economics in Guangzhou and Shenzhen, and we launched Europe's first commercial robotaxi service in Zagreb, Croatia. Our total cost per mile is the most competitive in the industry. The 2,000-vehicle commitment makes us Uber's largest autonomous driving partner in Europe, and we see substantial room to expand the fleet size further as performance and economics validate at scale.

Q: Could you please elaborate on your strategy going forward for the joint-deployment model? Can you share more color on how the commercialization model works and operates under an asset-light model? A: James Peng (CEO) and Leo Wang (CFO): The joint-deployment model accelerates fleet expansion with high capital efficiency. In this model, Pony supplies the Gen-7 robotaxi with Virtual Driver capability, a mobility platform introduces user demand, and an operating company handles fleet management. This creates a win-win alliance without disrupting existing ecosystems. Financially, the model generates sharing-based revenue or technology licensing fees, broadening the revenue base and introducing higher-margin recurring income. The 4,000-vehicle commitments from Uber and other partners will serve as a multi-year growth catalyst for 2026 and beyond.

Q: I have one question regarding PonyWorld 2.0. Can you please provide more color on what makes self-evolution different in autonomous driving, and how does it improve your R&D efficiency? If someone open sources a world model, would your models be affected? A: Tiancheng Lou (CTO): Autonomous driving is physical AI, and a general-purpose open-source world model is just a 3D video generatornowhere near enough to train our system. PonyWorld 2.0 is a self-evolving system that improves world model precision by matching exact real-world probability distributions of traffic participant behaviors, which vary from city to city. AI drives the whole process, with humans involved mostly for verification. This significantly reduces engineering resources needed to enter new cities, allowing us to enter many new markets simultaneously. This scaling ability is a large moat that won't be affected by open-source generative models.

Q: How should we think about Pony's new outlook for the domestic market heading into the second half of the year? A: James Peng (CEO): China is our home base, and domestic fleet expansion remains a significant part of our vehicle rollout. Our strategy focuses on the highest-valued markets firstTier 1 cities account for a significant share of national ride-hailing demand and offer the most mature regulatory frameworks. In Guangzhou and Shenzhen, expanding fleet size shortens wait times, boosts retention, and directly translates into higher daily revenue per vehicle. We will continue deploying more fleets in Tier 1 cities to widen our competitive moat, while also entering key Tier 2/3 cities such as Changsha, Hangzhou, and additional Greater Bay Area cities to establish new growth engines.

Q: Could you give us more color on how operational efficiency is being achievedfor example, on the remote assistant side, vehicle utilization, or charging and maintenanceand how these efficiency gains are helping accelerate deployment? A: Tiancheng Lou (CTO): Efficiency comes down to the fleet-to-staff ratio. Traditional taxis require a one-to-one ratio, but our robotaxis require zero human assistance when returning to depotsthey autonomously navigate, locate chargers, and self-park. We need only three people for every 100 robotaxis to keep daily operations running smoothly. This translates into significantly lower operating costs per vehicle and advanced unit economics. We've developed this know-how into standardized operating procedures and automation tools, which is why more partners are adopting our joint-deployment model.

Q: Could you give us updates on your new business initiatives, specifically the progress with your L4 light truck business? A: James Peng (CEO): The L4 light truck leverages robotaxi driving capabilities and cost-efficient hardware while sharing the same customer base as our robotruck business. It extends our logistics portfolio from long-haul into urban delivery, unlocking a new TAMChina has over 8 million active light trucks. Compared to low-speed robovans, our light truck offers three to four times the cargo capacity and is two times faster. The vehicle is jointly developed with CATL and is the world's first automotive-grade, fully redundant light truck built for L4. We've secured partnerships with SF Express and China Post Technology, with orders and deployment schedules already in place.

Q: We know that Waymo's management recently said that a demo is only 1% of the work. Could Pony's management share your views on this comment? A: Tiancheng Lou (CTO): This captures something realbuilding a demo and scaling are entirely different games. Autonomous driving is a probability problem. A typical ridesharing vehicle drives about 300 km/day, so a fleet of 100 cars generates tens of thousands of kilometers daily. If you get one accident every 1,000 km, that's 10 accidents per day at scaleno regulator or public would tolerate that. Going from demo to full scaling takes multiple 10x jumps in performance, and every jump is harder than the last. For Pony, we've already checked both boxesproven safety and rapid iterationwhich is why our focus is on expanding into more cities and deploying larger fleets.

Q: Congratulations on the strong quarterly results. Could you share more details on the financial performance and what drove the acceleration in robotaxi revenue? A: Leo Wang (CFO): Total revenues reached $36.2 million, a 69% increase year-over-year. Robotaxi revenue grew 691% to a record $12.1 million, with fare-charging revenue surging 849%. This acceleration was driven by fleet expansion into core downtown areas with high economic value and significant momentum from our joint-deployment model, including the successful launch in Zagreb, Croatia. Operating margin narrowed dramatically from -285.6% in Q2 2025 to -181.5% this quarter, and net loss narrowed 14.9% year-over-year. Our revenue growth significantly outpaced non-GAAP operating expense growth of just 9.6%, clearly demonstrating economies of scale and operating leverage.

For the complete transcript of the earnings call, please refer to the full earnings call transcript .