Pony AI Inc. (NASDAQ: PONY) announced the launch of PonyWorld 2.0, an upgraded world model for its autonomous driving technology that can diagnose its own weaknesses and guide targeted data collection.

The system includes three capabilities: self-diagnosis, targeted data collection in scenarios where the model underperforms, and training focused on difficult cases. PonyWorld 2.0 uses a structured intention layer that allows the model to form internal representations of its decisions, enabling large-scale self-diagnosis.

"PonyWorld 2.0 is an important step toward a more self-improving approach to autonomous driving development," said Dr. Tiancheng Lou, founder and CTO of Pony.ai. "As AI systems become more capable, they can play a larger role not only in learning to drive, but also in guiding their own improvement."

The company has been developing PonyWorld since 2020 as a reinforcement learning training system spanning cloud-side training and vehicle-side deployment. The upgraded system is being applied across Pony.ai's Level 4 driverless fleet to improve safety, ride comfort and traffic efficiency.

Pony.ai targets operating more than 3,000 vehicles by the end of this year, with deployments across 20 cities globally. The company states that nearly half of those cities will be in overseas markets. The autonomous driving company has validated unit economics of robotaxi operations in two metropolitan markets in China with its seventh-generation robotaxi fleet.

The system can review its own decisions, compare intent with outcomes, and identify scenarios requiring additional learning. It can generate targeted data-collection tasks for human teams to gather real-world samples that help recalibrate the world model for training.

Founded in 2016, Pony.ai operates across China, Europe, East Asia, the Middle East and other regions, according to the company's statement.