DeepSeek, the Chinese artificial intelligence laboratory that stunned Silicon Valley in early 2025 by matching leading US models at a fraction of the cost, is planning to raise up to 50 billion yuan ($7.35 billion) in what would be the largest funding round by a Chinese AI company.
The round would value the company at approximately $45 billion, up from $20 billion just weeks earlier, according to reports from The Information, the Financial Times and Bloomberg.
China's state-backed China Integrated Circuit Industry Investment Fund, commonly known as the Big Fund, is in talks to lead the financing, with cloud giants Tencent and Alibaba also in discussions to participate.
Liang Wenfeng, the billionaire hedge fund manager who founded DeepSeek and controls nearly 90% of the company, plans to write the largest cheque himself.
The fundraise marks a significant shift for a company that has until now operated more like a research laboratory than a commercial venture, funded entirely by Liang's quantitative hedge fund, High-Flyer, rather than by venture capital or strategic investors.
The decision to seek outside investment was driven partly by the need to offer equity to employees amid intensifying poaching of AI researchers by competitors including ByteDance, Alibaba and a growing cohort of well-funded Chinese AI startups such as MiniMax and Moonshot AI.
DeepSeek rose to global prominence in January 2025 after releasing a large language model that achieved performance comparable to OpenAI's GPT-4 while training on dramatically less computing power and at a fraction of the cost, a breakthrough that briefly wiped hundreds of billions of dollars from the market capitalisation of Nvidia and other US chip companies.
The company has since continued to close the gap with frontier US models in reasoning and coding benchmarks, and has kept its models open-weight, meaning versions are freely available for developers to download and modify.
The funding has also prompted DeepSeek to accelerate plans to generate revenue.
The company has told investors it intends to increase the frequency of model releases to align with industry norms, with an update to its latest model, DeepSeek V4.1, expected in June.
Related reading
- DeepSeek readies V4 model as Chinese AI challenger targets coding supremacy
- The $10 trillion race: How Nvidia and Alphabet are powering, and profiting from, the AI industrial revolution
- Sony deepens partnership with TSMC on AI image sensors
The scale of the round reflects a broader recalibration in China's AI sector, where the government is channelling state capital into domestic AI champions as competition with the United States intensifies and access to advanced Nvidia chips remains restricted by export controls.
For the global AI industry, DeepSeek's ability to attract $7 billion on a $45 billion valuation, while having generated minimal commercial revenue, underscores how aggressively investors are betting that the most capable AI laboratories will eventually capture enormous markets, regardless of which side of the US-China technology divide they sit on.