As the competition in AI large models moves into a new phase of "world models," how to enable AI not only to understand visuals but also to create truly usable 3D digital assets has become a new battleground for the industry. Recently, AI company Yingmou Technology announced a new round of financing worth hundreds of millions of yuan, led by Kaifu Fund and Shanghai Guotou Daliang, with existing shareholders continuing to invest. This is not only a recognition from the capital market of the 3D generation sector, but also marks that this Chinese R&D team mainly composed of "post-00s" individuals is beginning to shine on the international stage through solid technology.

The newly released next-generation model by Yingmou Technology Rodin Gen-2.5 represents a major technological milestone in the industry. It is not only the world's first 3D large model with the ability to generate tens of millions of faces, but it also introduces a "think first, then generate" mechanism similar to that of large language models. With this design, the model can adaptively handle objects of different complexities based on computational budget: simple models are quickly generated, while complex characters or detailed parts receive more computing power, achieving precise restoration of details such as skin texture, pores, and dragon scales.

image.png

Different from most generation tools in the market that can only achieve "presentation-level" effects, Yingmou Technology pursues "production-level usability." With the support of Rodin Gen-2.5, the generated models support physically accurate PBR materials at 12K ultra-high resolution, and the assets can be directly imported into game engines or post-production pipelines for disassembly, binding, and secondary editing. This practical orientation of "working well" has quickly won the favor of the overseas market, with about 80% of the company's revenue coming from North America, and its technology has been adopted in high-profile scenarios such as Unity AI Beta, Figma, and Huang Renxun's CES speech by NVIDIA's founder.

This breakthrough is driven by a group of young R&D talents. The core algorithm team of Yingmou Technology mostly comes from the MARS Lab at Shanghai Tech University, making it a truly "post-00s" team. In the context of high talent mobility in the AI industry, they have maintained a high level of talent stability, focusing on a vertical path from scanning the real world to generating digital humans, and then building complete 3D assets.