DeepSeek made its name building AI on a shoestring. Now it's raising $7 billion.
The Chinese lab is in talks for its first outside funding round since launch, a raise that would still leave it a fraction of OpenAI and Anthropic's recent funding rounds and valuations.
DeepSeek is in talks to raise about $7 billion in fresh capital. It would be the first time the lab has taken money from anyone other than its parent, the quant hedge fund High-Flyer.
For a company that built its name on doing AI cheap, the size is striking. DeepSeek's V3 model reportedly trained for about $6 million - a sliver of what US labs spend on a single training run.
The new round is roughly a thousand times that figure.
Reuters reports the round could value DeepSeek at $52 billion to $59 billion, with Tencent and battery giant CATL lined up as the largest outside backers and founder Liang Wenfeng committing roughly $3 billion of his own money.
Every weekday morning, Market Briefs breaks down what moves like this mean for your portfolio - in five minutes a day, with a free investing masterclass thrown in when you sign up.
DeepSeek doesn't have an obvious cash problem. High-Flyer has paid the bills from day one, and the lab has been more careful with compute than almost any rival.
The raise is about scale, not survival. Training the next wave of models takes more chips, more data centers, and more power than any single backer can keep funding alone.
Even OpenAI - which started with a billion-dollar check from Microsoft in 2019 - has gone back to investors round after round to keep pace.
DeepSeek rattled the AI world in January 2025 when its R1 model matched the reasoning quality of OpenAI's top systems at a fraction of the cost.
Markets noticed - Nvidia shed close to $600 billion in market value in a single day as investors rethought how much computing power frontier AI really needs.
Cheap AI was a story. Cheap AI with $7 billion behind it is a different story.
US officials have spent two years trying to choke off China's access to advanced chips. The bet was that compute limits would slow Beijing's AI plans.
DeepSeek already poked holes in that idea. A well-funded DeepSeek pokes more.
The raise also signals where China's AI playbook is heading. Until now, the country's leading models have mostly come out of internal labs at giants like Alibaba and Tencent, plus ByteDance's Doubao.
DeepSeek is the first standalone Chinese AI company raising at a level that puts it in direct conversation with Silicon Valley - even if its valuation still trails the top US labs.
Three things matter from here: who writes the checks, what valuation (how much the company is worth) gets agreed on, and whether Washington tightens chip export rules in response.
The valuation will tell investors whether global capital sees DeepSeek as a real challenger to US labs, while the investor list will reveal who's willing to bet on Chinese AI in the current political climate.
A $7 billion round changes the math on the AI race.
If you want this kind of read on the AI race every morning, join the investors reading Market Briefs - you also get a 45-minute investing course as a bonus.
