•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Pony AI Inc. (Pony.ai) is advancing the commercialization of Level 4 (L4) autonomous driving through continued cost reduction, fleet expansion and broader product deployment across mobility and logistics, the company said ahead of Auto China 2026 in Beijing.
At the event, Pony.ai highlighted progress in lowering the cost of its seventh-generation (Gen-7) Robotaxis. The company expects the total vehicle cost for its 2027 Robotaxi in the China market—including both the autonomous driving kit and the base vehicle—to fall below RMB 230,000. Pony.ai said this would be lower than the current starting price of a locally produced Tesla Model 3 in China.
Pony.ai said it has expanded its Robotaxi fleet from 270 vehicles to more than 1,400. The company attributed the progress to a 70% reduction in bill-of-materials cost for its Gen-7 autonomous driving kit, along with continued operational optimization.
The company also reported commercialization milestones, including unit-economics breakeven in two of Southern China’s largest economic hubs. Pony.ai said its registered user base has surpassed 1 million, nearly tripling from a year earlier.
Dr. James Peng, Founder and CEO of Pony.ai, said the focus has shifted from whether Robotaxi can work to how to scale it safely, efficiently and at the right cost.
Pony.ai said the latest cost reductions are expected to support mass production and further advance its “dual-engine” growth strategy across China and international markets.
The company aims to expand its fleet to more than 3,000 Robotaxis by the end of this year and extend its footprint to 20 cities globally, with nearly half of those cities in overseas markets. Based on its Gen-7 Robotaxi platform, Pony.ai plans to introduce an overseas version tailored to local regulatory requirements, infrastructure conditions and user needs, supporting deployment of more than 1,000 units internationally.
The company also cited collaboration with Toyota. It said the Gen-7 bZ4X Robotaxi has received an on-road testing permit in Guangzhou, and that as production ramps up, the two companies plan to deploy 1,000 bZ4X Robotaxis this year across major tier-one cities in China.
Pony.ai unveiled an L4 electric light-duty truck at Auto China 2026, described as the world’s first 100% automotive-grade, fully redundant light truck purpose-built for L4 autonomous driving and developed in collaboration with CATL.
Built on CATL’s Kunshi chassis platform, the vehicle is expected to reduce freight cost per kilometer by 40% to 50% compared with human-driven transport. Pony.ai said the truck offers 18 cubic meters of cargo capacity—2.6 times that of low-speed autonomous delivery vehicles—and is designed with a battery range of 320 to 450 kilometers for urban logistics scenarios including parcel distribution, supermarket restocking and cold-chain foodservice delivery.
The company said the truck is engineered for full-scenario and all-weather L4 autonomous driving, featuring full redundancy across steering, braking, communications, power supply, the computing platform and the sensor suite. Pony.ai also stated the vehicle is designed for a service life of 600,000 kilometers, or 20,000 operating hours.
Pony.ai said the launch strengthens its Robotruck portfolio, which already includes commercial Robotruck services across trunk-line logistics, dedicated-route logistics and port transportation.
The company also reported permits for Robotruck platooning operations on the Beijing-Tianjin Expressway and the Beijing-Tianjin-Tanggu Expressway. Pony.ai said a safety operator will be in the lead vehicle while following trucks operate without a human driver in the cab.
Pony.ai said the new L4 light-duty truck shares the same technology stack as its Gen-7 Robotaxis and is intended for deployment in the same operating regions. The company said this enables both platforms to leverage the same operational backbone, including energy replenishment, ground support, service centers and remote assistance, supporting economies of scale and improved marginal economics for large-scale deployment.
Dr. Tiancheng Lou, Founder and CTO of Pony.ai, said the company has completed a major upgrade of its core AI model, PonyWorld 2.0, described as the latest evolution of the training system behind Pony.ai’s autonomous driving stack.
Pony.ai said PonyWorld is not a conventional simulation tool, but a full reinforcement-learning system spanning cloud-side training and onboard deployment. As its fully driverless fleet has scaled, the company said improving autonomous driving performance has increasingly depended on improving the world model, particularly its ability to represent real-world dynamics and traffic interactions with sufficient accuracy and realism.
The company said PonyWorld 2.0 can diagnose where performance remains weak, guide more targeted data collection and support more efficient model training—creating a closed-loop data flywheel in which larger-scale driverless operations generate more valuable real-world data that improves the world model and strengthens the onboard system.
Pony.ai also emphasized safety as the foundation of autonomous driving commercialization. Lou said fail-operational capability across the entire system should become a universal industry standard for Level 4 autonomous driving, describing it as a way to maintain core driving functions through comprehensive redundancy and execute a safe pull-over maneuver in the event of unexpected hardware or software failure.
Premium gym chains are entering a “golden era” that is ending or already in decline, as rising operating costs collide with shifting consumer preferences toward more flexible, community-based ways to exercise. Long-term memberships are shrinking, margins are pressured by higher rents and facility expenses, and competition from smaller, more personalized…
