DeepSeek, the Chinese AI startup that challenged global assumptions about the country's ability to compete in advanced model development, is raising approximately 50 billion yuan ($7.4 billion) in its first external funding round at a valuation of up to 400 billion yuan ($59 billion).
Founder Liang Wenfeng has committed 20 billion yuan of his own capital to the round, according to sources familiar with the matter. Tencent Holdings is considering an investment of 10 billion yuan, while battery giant CATL is evaluating a 5 billion yuan stake. The final investor group is expected to include no more than 10 parties.
"This round signals that China's largest tech and industrial players are placing concentrated bets on domestic AI leadership rather than spreading capital across dozens of startups," a person with direct knowledge of the fundraising said.
The funding comes 18 months after DeepSeek's V3 and R1 models stunned the AI industry by matching or exceeding the performance of frontier models from OpenAI and Anthropic at a fraction of the reported training cost. DeepSeek's breakthrough — achieving competitive benchmark scores using fewer Nvidia H100 GPUs than US rivals — challenged the prevailing view that China's export restrictions on advanced semiconductors would cripple its AI ambitions.
Who Wins, Who Loses
Tencent's potential 10 billion yuan investment gives it a strategic foothold in foundational AI at a time when its own Hunyuan model competes with Alibaba's Qwen and ByteDance's Doubao. For Tencent, which reported 660 billion yuan in 2025 revenue, the 10 billion yuan commitment represents roughly 1.5 percent of annual sales — a calculated bet on owning the model layer rather than licensing it.
CATL's participation is more unusual. The world's largest EV battery maker, valued at over 1 trillion yuan on the Shenzhen exchange, has been expanding into AI data center power solutions and energy storage. Its 5 billion yuan investment suggests the company sees AI infrastructure as a natural extension of its battery and grid-scale storage business.
Alibaba and ByteDance, whose models compete directly with DeepSeek, are notably absent from the investor list. Their exclusion means DeepSeek will operate as an independent entity backed by Tencent's distribution ecosystem and CATL's industrial heft, rather than being absorbed into a single corporate AI strategy.
What the Valuation Means
A post-money valuation of 350 billion to 400 billion yuan would place DeepSeek among the most valuable AI companies globally. For context, OpenAI was valued at $300 billion in its most recent funding round, while Anthropic reached $61.5 billion. DeepSeek's valuation, at roughly one-fifth of OpenAI's, reflects both the discount applied to Chinese AI companies amid geopolitical uncertainty and the reality that DeepSeek has not yet built a significant revenue stream from its models.
The company has not disclosed plans for an initial public offering. Additional investors under consideration include Hong Kong-based IDG Capital and Monolith Capital, as well as China's national artificial intelligence fund, gaming company NetEase, and e-commerce leader JD.com.
The concurrent selling pressure on Hong Kong-listed shares of the involved companies suggests some investors view the capital deployment with caution. Tencent fell 3.7 percent on the day of the report, while JD.com dropped 4.2 percent and NetEase declined 1.4 percent. JD.com's short-selling ratio reached 46.2 percent, indicating significant bearish positioning.
For investors watching China's AI race, the key question is whether DeepSeek can translate its model performance into commercial revenue before the next generation of models from Alibaba or ByteDance closes the performance gap. The 50 billion yuan war chest gives it roughly 18 to 24 months of runway at current industry burn rates — enough time to prove whether its technology can generate sustainable returns.
This article is for informational purposes only and does not constitute investment advice.