DeepSeek has built one of China’s most formidable AI research operations without ever accepting a single yuan of external equity. Founded in 2023 as the AI research arm of High-Flyer, a quantitative hedge fund run by Liang Wenfeng, the company has operated on internal capital since its inception. That is now changing, but entirely on DeepSeek’s own terms.
According to reports from The Information, corroborated by Finance Biggo and Startup Fortune, DeepSeek is seeking to raise up to 50 billion yuan, or approximately $7.4 billion, in what would be the largest private funding round in the history of China’s AI industry. The round would value the company at 350 billion yuan, equivalent to $51.5 billion.
The decision to seek outside capital is not a sign of financial distress. Liang Wenfeng himself is expected to serve as the lead investor, committing approximately 20 billion yuan ($2.9 billion) of his own money. According to the Bloomberg Billionaires Index, Liang’s net worth is approximately $16.7 billion. The state-backed National Integrated Circuit Industry Investment Fund, commonly known as the Big Fund, is expected to be the second-largest investor in the round.
Why DeepSeek Is Raising Capital Now
The timing of the funding round reflects a convergence of pressures and opportunities. DeepSeek’s monthly active user base has reached 139 million, placing it behind ByteDance’s Doubao and Alibaba’s Tongyi Qianwen and Quark in the Chinese AI assistant market. Yet its website draws 355.5 million monthly visits, the highest of any Chinese AI platform, underscoring the scale of its reach even relative to better-capitalized rivals.
The company plans to use the proceeds to bolster its computing infrastructure, retain and recruit AI talent, and accelerate commercialization. On the talent front, the pressure is real: Luo Fuli, a contributor to DeepSeek’s V3 model, has left for Xiaomi, while Guo Daya has departed for ByteDance. Competing with companies for top researchers with far larger payrolls requires a war chest.
DeepSeek is also preparing to release V4.1 in June. The update is expected to add enterprise tools, support for the Model Context Protocol (MCP), and image and audio processing capabilities, though text will remain the only output format. Building and serving models of this scale demands substantial compute investment.
Alibaba and Tencent Both Turned Away
The most striking aspect of the funding round is not its size but who has been excluded from it. Alibaba, which had been in active negotiations with DeepSeek, was ultimately turned away. The core conflict was structural: Alibaba sought to integrate DeepSeek into its broader ecosystem, a strategy the company has pursued aggressively with its own Tongyi Qianwen model, embedding it across Taobao, AutoNavi, and Alipay through its Token Hub platform launched in March 2026. DeepSeek’s leadership viewed this as incompatible with the independent research culture that has defined the company.
Tencent’s approach was more direct. According to sources, Tencent requested a 20% stake in DeepSeek. That request was also rejected.
The outcome is a funding round structured to minimize strings attached. As Jiang Yi, Managing Partner at Hengye Capital Management, described it, the round is designed to be “the offer with the fewest strings attached.” The Big Fund’s expected participation is notable precisely because state-backed capital of this kind typically does not come with demands for ecosystem integration or board control.
Independence As a Strategic Asset
DeepSeek’s insistence on independence is not merely a cultural preference, it is a competitive strategy. As China’s AI market consolidates around a handful of dominant platforms, an independent DeepSeek occupies a unique position: it can collaborate with, or compete against, any of the major tech conglomerates without being subordinate to any of them.
This dynamic has become increasingly important as China’s AI ecosystem fragments into what some analysts have described as a multi-faction competitive landscape. ByteDance, Alibaba, Tencent, Huawei, and Baidu are each building proprietary AI stacks, and each has an interest in controlling the models that run on their infrastructure. DeepSeek’s refusal to be absorbed by any of them preserves its ability to operate across all of these ecosystems simultaneously.
The company’s open-weight model releases have also given it a kind of structural influence that no single corporate patron could replicate. By releasing models that developers worldwide can download, fine-tune, and deploy, DeepSeek has built a global developer community that is independent of any single platform’s distribution channel. As EastFrontier has previously reported, this approach has made DeepSeek one of the most consequential forces in the global open-source AI ecosystem. (See our coverage of China’s AI ecosystem entering a new competitive phase and DeepSeek’s role in the US-China AI rivalry.)
What the Round Signals for China’s AI Investment Landscape
DeepSeek’s funding round is a landmark not just for the company but for the broader Chinese AI investment market. It demonstrates that a Chinese AI startup can command a $51.5 billion valuation without being absorbed into one of the country’s tech conglomerates, and that there is sufficient capital available, including from state-backed sources, to support that independence.
It also signals the market’s maturation. In the early years of China’s AI boom, the dominant model was for startups to seek strategic investment from Alibaba, Tencent, or Baidu to gain distribution and resources. DeepSeek’s round suggests that the most capable AI companies now have enough leverage to set their own terms.
For investors, the round presents a rare opportunity to back a company that has demonstrated both technical excellence and commercial traction without the constraints of ecosystem dependency. Whether the round closes on the terms reported will be closely watched by everyone in China’s AI industry.