Jingshuo, a subsidiary of WeRide that develops data infrastructure for physical AI, has secured over US$10m in investment as part of its Pre-Series A funding round, with investors including Vision Knight Capital, Hong Kong Science and Technology Parks Corporation (HKSTP) and Sirius Capital.
Jingshuo is building data and model infrastructure for physical AI, guided by the principle of “driving Physical AI evolution through high-quality data.” For autonomous vehicles, Jingshuo’s data infrastructure could support the training of AI models using diverse real-world driving scenarios, helping autonomous systems interpret their surroundings, predict the behavior of other road users and make driving decisions in complex environments.
Jingshuo has developed a product portfolio built on its integrated data-and-model platform WorldEngine and proprietary generative world model GENESIS-Robotics that spans EGOK, its high-precision first-person data collection device; large-scale data acquisition; high-quality datasets; and customized data services. Together, these technologies link real-world operational data with robot learning.
As artificial intelligence expands beyond the digital world into physical environments, high-quality data and the infrastructure supporting it are becoming increasingly important to technological and industrial development. The capabilities of large language models (LLMs), world models and embodied AI systems depend on the continuous collection, management and refinement of high-quality data. This is particularly important in Physical AI, where the collection, processing, validation and use of real-world training data help determine the capabilities and limitations of AI models.
As part of WeRide’s business, Jingshuo builds on nearly a decade of WeRide’s experience in commercializing physical AI technologies. The company is further advancing its capabilities in world models, data platforms and its skill asset engine to build an integrated infrastructure spanning data production, data governance, model training and the development of reusable AI capabilities, providing foundation for the next generation of physical AI systems.
Following this funding round, Jingshuo plans to expand its core R&D team; strengthen the integration of its data platform, world model and skill asset engine; and broaden the task and scenario coverage of its high-quality datasets. These investments will advance Jingshuo’s efforts to build the data and model infrastructure needed to train, develop and scale next-generation physical AI systems.
Related news, Tier IV and TUM integrate Nvidia Hyperion to accelerate Autoware-based L4 AV projects in Germany
