Guanglun Intelligent raises ¥10B to scale embodied AI data and testing infrastructure

Guanglun Intelligent sits in a critical but often overlooked layer of the AI stack: high-quality embodied data, simulations, and evaluation infrastructure for robots and physical-world agents. A ¥10B strategic round anchored by state-linked funds and major industrial players signals that China is willing to treat this data-and-testing layer as national infrastructure, not just a startup niche. In practice, Guanglun is building the “data factories” that let robot and world‑model systems iterate safely before they ever touch the real world.

Strategically, that matters because embodied AI has historically been bottlenecked less by model architectures and more by trustworthy, diverse interaction data and robust sim‑to‑real pipelines. If Guanglun and its peers can mass‑produce high‑fidelity synthetic and mixed reality datasets, Chinese robotics and autonomous systems companies can move much faster without bearing those fixed costs individually. This is analogous to the way shared cloud infrastructure unlocked web startups a decade ago—only now it’s for robots and agentic systems.

For the race to AGI, heavy investment into embodied data infrastructure strengthens the path where advanced agents are grounded in physical environments, not just text and images. That doesn’t directly create AGI, but it accelerates the feedback loop between powerful models and messy real‑world tasks, which is exactly where general capabilities—and safety challenges—tend to sharpen.