Feijiekes Intelligent Technology

Fysics AI, legally known as Feijiekesi Intelligent Technology (Shanghai) Co., Ltd., is a core infrastructure provider in the Physical AI sector, established in August 2024. The company focuses on developing key technologies for embodied intelligence, which aims to transition artificial intelligence from digital interfaces to active engagement with the real world.

The firm's core product is a proprietary, next-generation universal physics simulation engine. This engine is designed to be differentiable, handle multiple physics and materials, and is a foundational technology for the entire lifecycle of robotics, including research and development, training, and deployment. A key feature is its ability to generate high-quality synthetic data for large-scale simulation training, addressing the scarcity of real-world data for training embodied AI. This process significantly lowers the costs and increases the efficiency associated with physical robot and environment setups. The engine is compatible with major GPUs, including those developed in China, and supports large-scale parallel simulations.

Fysics AI recently secured a Pre-A1 funding round of nearly 100 million RMB in January 2026. The round was co-led by Matrix Partners China and Oriental Fortune Capital, with participation from Muxi Integrated Circuit, Chixing Venture Capital, and others. The capital is allocated to refining the physics engine, developing foundational models for physical intelligence, creating data asset generation models, and expanding the commercial application of its products. The company's chairman is Zhang Lihua, who has been recognized as a "Shanghai Overseas Chinese Community Outstanding Figure." The CTO is Dr. Yang Dingkang. In collaboration with the Cognitive and Intelligent Technology Lab (CITLab) at Fudan University, Fysics AI launched FysicsWorld, a unified, full-modality evaluation benchmark for the real physical world.

Keywords: Physical AI, embodied intelligence, physics simulation engine, robotics, synthetic data generation, machine learning, artificial intelligence, FysicsWorld, GPU computing, parallel simulation, robotics training, digital twin, autonomous systems, AI infrastructure, venture capital, Matrix Partners, Zhang Lihua, Yang Dingkang