DISCOVER Robotics Raises $100M Angel+ Round: Embodied AI Now Valued as a Full Model-Data-Simulation Stack
Category: industry · Embodied AI / Funding Date: 2026-08-03 08:35 (Beijing Time) Sources: Leiphone exclusive report; DISCOVER Robotics official GitHub and project materials
The Numbers, First
Per Leiphone's report, DISCOVER Robotics completed a $100 million angel+ funding round on August 3. Investors include IDG, Xinglian, Wu Yuefeng, Dachen, Joyoung, Oriza Hua Ying, and the Binhu District industrial investment group, with existing shareholders also increasing their stakes. The report notes the company's earlier angel round of over $100 million closed less than a month prior.
An important caveat: while the article uses the phrase "embodied unicorn," no post-money valuation, equity split, or investment agreement has been publicly disclosed. A $100 million raise is not a $100 million valuation, and it certainly cannot be treated as an independently audited valuation figure.
It's Not Selling a Robot
DISCOVER Robotics splits its product roadmap into three layers:
- ORION: A hierarchical embodied intelligence foundation model handling higher-level perception, decision-making, and action generation.
- LOOP: A hybrid virtual-real data flywheel designed to connect simulation, real-robot feedback, and training data.
- DISCOVERSE: An open-source, high-fidelity robotics simulation framework built on MuJoCo and 3D Gaussian Splatting, covering workflows for data collection, imitation learning, diffusion policies, and RDT.
- https://www.163.com/dy/article/L3D54AQV05118HA4.html
- https://github.com/DISCOVER-Robotics/DISCOVERSE
- https://github.com/DISCOVER-Robotics
- https://air-discoverse.github.io/
The DISCOVERSE GitHub repository confirms it is an MIT-licensed, modular simulation system oriented toward Real2Sim2Real, supporting ACT, Diffusion Policy, RDT, and custom algorithms. However, the repository also explicitly states that the public README provides no unified comparisons of simulation speed or real-world task success rates. In other words, one can confirm that "the foundation exists and the interfaces are open" — but not that ORION's generalization capability has reached commercial delivery level.
Why It Matters
The signal from this round is not merely "yet another robot company raised big money." It's that capital is starting to pay for a complete closed loop: foundation models solve generalization, simulation solves data and long-tail scenarios, and real-robot feedback closes the reality gap. In embodied AI, the most expensive part is often not training a model once, but continuously acquiring enough clean, reproducible interaction data.
If this closed loop truly works, the company's moat will shift from robot hardware design to the compounding effects of "data generation → policy training → real-robot recycling." If it doesn't, the three-layer product could easily become three disconnected projects — and the funding number would remain just a funding number.
Original article and evidence: