
Kaiwang Data (恺望数据), a Chinese provider of data infrastructure for embodied AI, has completed a stra...
The AMW Read
Updates player map in embodied AI with a data infrastructure niche; segment-level significance for robotics data needs.
Kaiwang Data (恺望数据), a Chinese provider of data infrastructure for embodied AI, has completed a strategic funding round of more than RMB100 million (~$14M). The round was jointly led by the Beijing E-Town Industrial Upgrade Fund, Huafang Capital, and Skyline Capital (天际资本), with participation from robotics companies Lumai Robotics, Songyan Power, Zizibianliang Robotics, and AGIBOT-affiliated Mifeng Technology.
The company collects, labels, and manages multimodal data for autonomous driving, world models, and humanoid robots, covering first-person bare-hand video, tactile-glove data, gripper operations, and panoramic footage. Kaiwang Data reports monthly usable data output has stabilized at about 100,000 hours and has secured bulk data-purchasing agreements with several major companies. The new funding will support data-trading platform development, world-model technology development, and capacity expansion across commercial, industrial, and household settings.
Why it matters: This round underscores the critical role of data infrastructure in the embodied-AI value chain, a segment where high-quality, task-specific data is a key bottleneck. The participation of robotics companies as investors signals a strategic move to secure data supply chains, reflecting a broader pattern of vertical integration in China's robotics ecosystem. As humanoid robots and world models advance, the demand for diverse, high-quality embodied data is surging, positioning data infrastructure companies as essential enablers. This development also highlights the growing capital flow into data infrastructure within the robotics sector, distinct from the compute-centric focus of AI infrastructure more broadly.
Grounded expert take: Kaiwang Data's success reflects the increasing recognition that data, not just compute, is a decisive moat in embodied AI. The company's focus on tactile and multimodal data, coupled with stable bulk agreements, suggests it is building a defensible niche. The involvement of robotics firms as investors further validates the strategic importance of data access, potentially accelerating the commercialization of embodied-AI applications. However, the relatively modest round size and China-specific market dynamics may limit its immediate global impact, though it reinforces the centrality of data infrastructure in the broader AI substrate.