Funding news · Sep 16, 2026
Apex Intelligence Raises ¥400M Seed for Self-Evolving Foundation Models
Apex Intelligence Lands ¥400M Seed Round
Apex Intelligence, a Beijing-based foundation-model research startup, has raised ¥400 million in Seed funding. XtalPi led the round, with participation from a broad syndicate of investors.
What Apex Intelligence Does
Apex Intelligence develops self-evolving foundation models. The company's stated ambition is to automate the full research cycle — generating ideas, running experiments, and writing papers — with each research trajectory recycled as training data for the next model iteration. In other words, the models are designed to improve themselves by learning from the research they produce.
The company was founded in June 2026 by Chen Yongchao, an assistant professor at Tsinghua University's Institute for AI. Chen previously conducted research at Google Research, Google DeepMind, Microsoft Research, and the MIT-IBM Watson AI Lab.
Within roughly three months of founding, Apex Intelligence had grown to approximately 30 employees. The company also co-developed the ASI-Bench benchmark with Tsinghua, MIT, Harvard, CMU, USTC, and Microsoft Research — a notable signal of early research collaboration with leading academic and industry institutions.
Funding Details
| Detail | Value |
|---|---|
| Round type | Seed |
| Amount | ¥400,000,000 CNY |
| Date | September 16, 2026 |
| Lead investor | XtalPi |
| Investors | Chuxin Capital, XtalPi, Zhongguancun Science City, IDG Capital, Shenzhen Capital Group, Yunxiu Capital, Shanghai Future Industry Fund, Xinglian Capital (星连资本), SEE Fund, Dexun Capital (德迅资本) |
The investor list spans venture capital, government-linked funds, and strategic backers. XtalPi, an AI-driven drug discovery and materials science company, led the round. Zhongguancun Science City and the Shanghai Future Industry Fund represent institutional and regional backing, while IDG Capital, Shenzhen Capital Group, Chuxin Capital, Yunxiu Capital, Xinglian Capital, SEE Fund, and Dexun Capital round out the syndicate.
Why This Matters
The round stands out for its size at the Seed stage. A ¥400 million Seed round is unusually large and reflects both the capital intensity of foundation-model research and investor appetite for teams pursuing ambitious AI research automation.
The "self-evolving" framing is central to the company's pitch. Rather than building a static model, Apex Intelligence aims to create a system where research outputs feed back into model training — a closed loop that could, in theory, compound capability over time. The co-development of ASI-Bench with a roster of top-tier institutions suggests the company is investing in rigorous evaluation of that thesis from the start.
What to Watch
The key questions for Apex Intelligence will be whether its self-evolving approach produces measurable capability gains, and how it converts research momentum into durable commercial or scientific advantage. The involvement of XtalPi as lead investor may hint at application areas in scientific discovery, though the company's stated focus remains on automating the research cycle itself.
With roughly 30 employees and a substantial Seed war chest, Apex Intelligence has the resources to scale its research team and continue developing both its models and its benchmark work. The company's ability to attract top research talent — and to publish credible results — will be central to its next phase.
About Apex Intelligence
Apex Intelligence is a Beijing-based foundation-model research startup founded in June 2026 by Chen Yongchao. The company develops self-evolving foundation models aimed at automating the full research cycle, and co-developed the ASI-Bench benchmark with Tsinghua, MIT, Harvard, CMU, USTC, and Microsoft Research.