Nearly 200 million yuan! JD.com led the angel round financing for embodied intelligence company RoboScience

Recently, RoboScience, an embodied intelligence company, completed nearly 200 million yuan in angel round financing, led by JD.com, with participation from China Merchants Venture and SenseTime Sense Capital. Existing shareholder 01VC continued to invest, with Rollstone Capital serving as exclusive financial advisor for this round.

Established in December 2024, RoboScience officially began operations in March of this year,  focusing on developing world-leading embodied intelligence models and products for both large and small brains. Prior to this round, it completed a seed round of financing of tens of millions of yuan, exclusively invested by 01VC.

In addition, RoboScience has established a system for scalably acquiring and using simulated data, video data, and real-world data, covering a wide range of objects and tasks found in daily life, with varying dimensions, which improves data quality and reduces data costs.

In terms of technical route, RoboScience has adopted a fast and slow brain layered end-to-end model since its establishment.

Information indicates that RoboScience independently developed the Manipulation Foundation Model, which serves as a bridge between multimodal models and the physical world, achieving generalization across three dimensions: guiding any robot, manipulating any object, and completing any task.

For grasping, for example, compared to existing methods that rely solely on specific objects and specific robotic arms, this model can be decoupled from hardware and significantly improves success rate, posture diversity, and computational speed, providing a new solution for dexterous grasping.

Leveraging highly generalized technical capabilities, the company’s products can efficiently adapt to different hardware configurations based on scenario requirements, flexibly expanding their application scenarios. Cross-embodied AI, a core technology module, complements an embodied operating system developed based on a fast-slow brain system robotic learning framework and self-supervised training, forming the product’s intelligent core.

In terms of application scenarios, the company is providing solutions across multiple sectors, including industry, logistics, consumer retail, and home.