Fundraise

NeuxMind.AI raises $30M Series B to sell enterprise AI "workers

What's the deal? Beijing-based NeuxMind.AI has raised US$27.5M in a Series B round led by Saize Capital, with existing US dollar and renminbi shareholders participating. The company builds an enterprise agent operating system that dispatches digital and physical AI "workers" for large businesses.

What's the endgame? NeuxMind sells outcomes, not software seats. Its "Shenshou" Agent OS trains and runs AI workers under a results-as-a-service model, with pricing written into contracts and tied to revenue growth, cost cuts, or efficiency gains.

What does it serve? The company targets banks, financial institutions, and government bodies, with the Middle East — Saudi Arabia, Qatar, and the UAE — as its priority market. Founded in 2020 and led by Zhang Lei, it also serves Southeast Asia, Japan, and Greater China.

Why now? NeuxMind says revenue roughly doubled to about US$27.5M this year, and founder Zhang expects annual doubling to continue. The company reports more than US$13.8M in overseas revenue and says it has had no failed outcome-based deliveries so far.

What's next? NeuxMind plans to expand into physical AI, acquiring a robotics team to build a system that directs drones, robot dogs, and robotic arms. It aims to supply the "brain" for intent recognition and dispatch rather than the hardware itself, with products expected in the first half of 2027.

What could go wrong? The outcome-based model depends on results being measurable; Zhang warns that charging for fuzzy outcomes is "digging yourself a hole" (translated from Chinese). The company also faces crowded competition from enterprise agent platforms in China and global vendors such as Sierra.

The signal: China's enterprise-agent market is drawing heavy capital, but the harder test is whether deployments deliver returns. A Frost & SullivanDealroom has a profile for this one. Try Dealroom → survey found over 80% of enterprises use AI, while fewer than 40% say it has improved profits — the gap NeuxMind's pitch targets, even if the company-reported figures don't yet prove the model works.

Read more: chinaventure.com.cn

Image credit: "Quantum Computing for Google Goggles" by jurvetson, licensed under CC BY 2.0

More top stories