Purdy Ho

Purdy Ho asked for Pony AI's perspective on the surge of new entrants, particularly EV makers, into the Level 4 autonomous driving space, and the main technical and operational challenges they might face.

Answer

James Peng, Chairman and CEO, welcomed new entrants as a sign of increasing recognition and confidence in robotaxi commercialization, but emphasized the difficulty of entry, noting that no new players have fully driverless vehicles deployed. He outlined three major hurdles: business challenges (user acquisition, production, fleet management, early mover advantages), regulatory challenges (high safety requirements, lengthy licensing processes), and technical challenges. Tiancheng Lou, CTO, elaborated on the technical side, highlighting Pony AI's five-year head start in world model development (reinforcement learning, simulation) and its closed-loop training system that reduces reliance on real-world data. He explained how virtual driver feedback improves the world model's ability to handle corner cases and noted that the Hong Kong IPO will accelerate R&D.

Purdy Ho asked for Pony AI's perspective on the surge of diverse players, particularly EV makers, attempting to enter Level 4 autonomous driving operations, and the main technical and operational challenges for these new entrants, such as tackling corner cases and fleet management.

Answer

Chairman and CEO James Peng welcomed new entrants as a validation of the robotaxi industry's potential but identified three major hurdles: business challenges (user acquisition, production, dispatching, maintenance, early mover advantage, partnerships), regulatory challenges (high safety requirements, lengthy step-by-step proving processes, licensing), and technical challenges. CTO Tiancheng Lou elaborated on the technical side, emphasizing Pony AI's five-year head start in world model development (reinforcement learning based on simulation), which enables unsupervised, self-improving iterations and reduces reliance on real-world data. He explained how the world model and virtual driver co-evolve to handle corner cases and stressed Pony AI's advantage in validating and deploying new technology at scale, further reinforced by the Hong Kong IPO.