What happened
Shanghai Huan Yu Xin Sheng Technology Co., Ltd., known as Huan Yu Xin Sheng, has completed a seed funding round of tens of millions of yuan, with investment from Yuan Sheng Capital.
The funds will be allocated to core technology R&D, world model training infrastructure, multimodal data collection systems, and team expansion.
Founded in 2026 and headquartered in Pudong New Area, Shanghai, the company focuses on neuro-cognitive world models, integrating embodied intelligence and brain science.
Why it matters
The company's self-developed neuro-cognitive world model, featuring a 'memory-knowledge-inference-calibration' loop and a brain-inspired fast-slow planning architecture, has shown strong performance in the World Arena 2.0 benchmark, ranking top three in Track 1 and second in Track 2.
This suggests a potential breakthrough in small-sample learning, training efficiency, and real-world generalization, moving beyond traditional state predictors that rely on parameter scaling.
The founding team, led by CEO Ren Yunfan, a 1999-born PhD from the University of Hong Kong and postdoc at the University of Zurich, brings expertise from top institutions and companies like Microsoft, Tesla, and Meta, positioning the startup at the intersection of AI, robotics, and neuroscience.
Key facts
Huan Yu Xin Sheng completed a seed round of tens of millions of yuan, with Yuan Sheng Capital as the investor.
The company was founded in 2026 and is based in Shanghai's Pudong New Area.
In the World Arena 2.0 evaluation at IROS 2026, the company ranked top three in Track 1 and second in Track 2.
CEO Ren Yunfan is a 26-year-old national-level young talent and a tenured associate professor at Tongji University.
What to watch next
How the company leverages the new funding to scale its world model training infrastructure and data collection capabilities.
Whether its neuro-cognitive approach can maintain a competitive edge in future benchmarks and translate into practical embodied AI applications.
Potential collaborations with academic institutions like Tongji University's Spatial Intelligence Lab, given the team's ties.