Deepseek went looking for about $7.5 billion. It’s now close to raising at least $12 billion, Bloomberg reported, and people familiar with the matter said the final total could approach $15 billion.
That’s a lot of extra money in one round. The original target already came with a valuation of roughly $75 billion.
A battery giant and a gaming giant write the biggest checks
Two of the names putting in money aren’t surprising. Battery maker CATL and Tencent are contributing the largest shares of the round.
None of the three companies would talk about it. CATL declined to comment. Tencent and Deepseek didn’t respond.
Investors are interested because of one model. Deepseek’s V4-Flash has set new benchmarks for cost and performance against Anthropic and OpenAI. But keep in mind what investors are paying for: Deepseek released V4-Pro and V4-Flash in April as preview versions, not finished products.
The round almost didn’t happen
The deal stalled earlier this year. Deepseek paused the round after comments by founder Liang Wenfeng about the company’s reliance on Nvidia chips went viral.
The jump in valuation over just a few months is easy to see. In June, Deepseek closed its first external funding round at about $7.4 billion, with a valuation above $50 billion. This new round could bring in roughly twice that amount.
Huawei chips now, its own chips later
Deepseek is clearly trying to put the Nvidia question behind it. The company is building a data center with at least 160,000 AI chips from Huawei.
Huawei isn’t the long-term plan either. Deepseek is also working on its own inference chip to reduce its dependence on both Nvidia and Huawei.
Next stop, the public markets
After the round closes, Deepseek plans to restructure for an IPO in early 2027.
Liang wants Deepseek to keep developing open models with freely available weights. If you’ve downloaded a Deepseek model, that’s the commitment to watch. Liang is promising to keep giving the core product away for free while asking investors for as much as $15 billion.
