Pony.ai has launched public road testing of its seventh-generation robotaxis – jointly developed with GAC – in China, ramping up its robotaxi mass production and commercialization efforts.
Vehicles currently being tested on public roads in the southern Chinese cities of Guangzhou and Shenzhen are based on GAC’s second-generation Aion V model and are equipped with Pony.ai’s latest Gen-7 autonomous driving system.
Pony.ai’s Gen-7 autonomous driving system
Pony.ai’s Gen-7 autonomous driving system features three advances: the exclusive use of 100% automotive-grade components, a 70% reduction in BOM costs of the autonomous driving kit compared with the previous generation, and an enhanced platform-based design enabling rapid adaptation across multiple vehicle models.
Equipped with six types of mass-production sensors providing 360°, blind-spot-free redundant perception, and incorporating Pony.ai’s proprietary sensor cleaning solution, the system is said to effectively handle adverse weather and extreme scenarios. It is built on Pony.ai’s in-house PonyWorld model and uses reinforcement learning to support safe operation. The system is designed to navigate complex, real-world conditions across a range of scenarios and weather conditions, with the goal of continuous performance improvement.
Dr James Peng, CEO and co-founder of Pony.ai, said, “The Aion V with 7th-Gen AD system deployed in these road tests represents a significant milestone in our deep collaboration and ecosystem synergy with GAC. Since launching our strategic alliance in 2018, we have continuously deepened our cooperation across L4 autonomous vehicle R&D, product deployment and strategic investments.”
2025 mass-production
Pony.ai is aiming for mass production in 2025 as it moves from technology demonstration to large-scale deployment. With improvements in cost-efficiency in its latest Gen-7 system and readiness for pre-installed production, the company plans to expand its fleet to over 1,000 vehicles by the end of the year, supporting greater scalability.