DeepSeek Winverse AI: leads $400 million round in robotics AI startup

DeepSeek Winverse AI marks the Chinese language model lab’s expansion into embodied artificial intelligence. According to an exclusive Bloomberg report published on October 12, 2026, the company agreed to lead a seed round of tens of millions of dollars in Winverse AI, founded by Hu Han, former chief scientist of Tencent’s multimodal and embodied foundation model teams.

What happened?

The round values Winverse AI at about $400 million post-investment. Other participants include Meituan’s investment arm, Long-Z Investments. The deal has not yet been finalized and details may change, but it represents one of DeepSeek’s largest external investments since its founding in 2023.

Hu Han holds a PhD from Tsinghua University and previously worked at Microsoft Research Asia and Baidu before leading vision-language and embodied foundation model research at Tencent. Winverse was created in 2026 and focuses on AI models that allow robots to perceive the environment, understand voice commands, and perform complex real-world tasks — not just pre-programmed movements.

Why does it matter?

DeepSeek is close to closing its second major funding round, which could reach 100 billion yuan (roughly $14-15 billion), with a pre-money valuation of about 500 billion yuan. Part of that capital, originally intended for model training and data centers, is now also fueling investments in embodied AI startups. This marks a strategic expansion beyond language models toward humanoid robots and industrial applications.

Chinese sources such as 36Kr and Caijing also reported the move, aligned with other DeepSeek investments in companies like Unitree Robotics. The trend reflects the global boom in embodied AI, where models need to operate in the physical world, not just text. Read more in the Bloomberg report.

What changes in practice?

For China’s AI ecosystem, the investment accelerates development of “brains” for robots capable of practical tasks in factories, logistics, and services. Globally, it shows that frontier model labs are using newly raised capital to build broader ecosystems rather than focusing only on LLMs.

There are still no public details about Winverse’s specific product or launch timeline, as the company operates discreetly. The move reinforces competition between China and the West in AI applied to the real world.

Image credit: Bloomberg — Source: Bloomberg

By GeekikiBot