DeepSeek has closed its first outside funding round, raising more than $7.4 billion in a deal that values the Chinese AI lab at up to $59 billion. The round marks the first time the company has accepted money from anyone outside High-Flyer, the quantitative hedge fund founded by DeepSeek’s own founder, Liang Wenfeng.
The round took in roughly 50 billion yuan and places DeepSeek among the largest privately funded technology companies in China. It also ends, at least for now, the lab’s reputation for staying away from outside capital even as rival AI companies chase larger checks.
An Unusual Way to Take Money
Most funding rounds give investors direct equity and voting power in the company they back, but DeepSeek’s deal works differently. Investors routed their capital into a limited partnership controlled by Liang rather than buying shares directly in DeepSeek itself. Most of those investors accepted a five year lock up and gave up voting rights as part of the arrangement.
However, one investor stood apart from the group. China’s National Artificial Intelligence Industry Investment Fund put its money straight into DeepSeek and kept its voting rights, a structure not extended to the other backers.
Liang himself contributed 20 billion yuan, close to 40% of the total raise. That large personal stake let him absorb the dilution that comes with outside money while keeping firm control over a lab that has long operated more like a research group than a startup chasing growth at any cost.
Tencent and CATL Lead the Backers
Tencent put in about 10 billion yuan and battery maker CATL contributed roughly 5 billion yuan, making the two companies DeepSeek’s largest outside backers. Other reported participants include gaming company NetEase, e-commerce giant JD.com, and Hong Kong based IDG Capital, alongside several state backed investment funds.
The mix of a social media giant, a battery manufacturer, and state capital reads like a coordinated push behind a company Beijing sees as central to its AI ambitions.
Still Far Behind the U.S. Pace
Even with the fresh capital, DeepSeek’s war chest looks modest next to its American competitors. The raise is still behind the $122 billion OpenAI gathered this year and the $65 billion Anthropic locked in last month.
Export restrictions that block DeepSeek from frontier American chips help explain that gap. But DeepSeek built its name on doing more with less. Its R1 model wiped close to $600 billion off Nvidia’s market value in a single trading session back in January 2025, simply by matching the performance of far costlier American AI models such as ChatGPT.
With $7.4 billion in fresh capital now behind it, the company that made frugality its brand now has the resources to fund the kind of expensive training its next generation of models will likely demand.
