DeepSeek is in advanced discussions for a funding deal that could raise $10 billion and value the company at $45 billion, Bloomberg reported Friday (May 22).
If DeepSeek raises $10 billion, it could set a record for a first-time financing by a Chinese tech startup, according to the report.
The discussions are ongoing, and the details could change, per the report.
The artificial intelligence startup’s founder, Liang Wenfeng, has told potential investors that DeepSeek is focused on pushing the boundaries of AI and pursuing the goal of artificial general intelligence rather than short-term monetization, according to the report.
That focus is “striking” at a time when other global AI leaders are exploring new ways to earn revenue, the report said.
“China is leveraging low power costs and a huge developer pool to treat AI tokens—basic data units processed by AI—as tradable assets,” Bloomberg Intelligence analysts Shirley Wong and Robert Lea said in the report. “The ‘industrial’ approach to AI is being fueled by a surge in one-person firms using tokens to raise productivity. Asia AI software stocks climbed 11.3% this year, outpacing the U.S.”
Advertisement: Scroll to Continue
The financing figures cited by Bloomberg are higher than those reported in April. The Information reported April 22 that DeepSeek hoped to raise funding at a valuation of more than $20 billion, following strong interest among investors. The publication had reported earlier in April that DeepSeek had begun talks to raise its first outside capital and hoped to bring in at least $300 million at a valuation of at least $10 billion.
That report said that DeepSeek’s focus on offering open-source technology over making money had led to some debate about how the company should be valued. DeepSeek hasn’t generated much revenue, as its models are open-source and its chatbot is free for consumers to use.
It was reported April 27 that DeepSeek had begun offering developers a temporary 75% discount on the use of one of its AI models amid increasing competition in China’s AI space.
The report said the move threatened to restart a price war that began when DeepSeek unveiled its R1 model in 2025 and said it had trained the model at a fraction of the cost paid by its larger rivals.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.