DeepSeek is preparing for a public offering. China’s artificial intelligence lab has begun laying the groundwork for an initial public offering. According to Bloomberg. He could file as soon as this year, with a debut slated for 2027. First, he wants to earn more privately.
The Hangzhou lab closed its first overseas tour just weeks ago. That increase took in about $7 billion, valuing it at about $50 billion. Now he’s talking about a new round with new backers. The initial estimate in the table is at least 480 billion yuan, or about $71 billion.
Bloomberg says DeepSeek is seeking at least 10 billion yuan, about $1.5 billion. The final figure may be several times higher depending on how many investors are involved. The Financial Times was the first to report on the new fundraiser. Information It is aiming even higher in a second round of around $7.4 billion.
Seven times in three months price increase again
The $71 billion tag would see a nearly 40% jump in six weeks. Move away and the climb is steeper. DeepSeek opened to foreign capital worth about $10 billion in April. Within a few days, Tencent and Alibaba entered into negotiations, and this figure exceeded 20 billion dollars. By May, it was close to $45 billion. In June, the round was closed at close to $50 billion.
The first growth was unusual in design. Only one investor, China National Artificial Intelligence Industry Investment Fundreceived direct capital and voting rights. Every other backer’s money went to a limited partnership controlled by founder Liang Wenfeng. They did not vote and were sentenced to five years in prison. The structure usually occupied one room. Instead, the round is reportedly oversubscribed.
IPO chart
DeepSeek is working with accounting and banking advisers to complete its financial statements by the end of December, one of the people told Bloomberg. This is a required step before any documentation. The listing could be late this year or early 2027. The time depends on when the figures are ready.
No exchange confirmed. The market awaits mainland China or Hong Kong, where local rivals Zhipu and MiniMax are already listed. Zhipu’s stock has risen nearly 1,600% since its January debut. This precedent helps explain why DeepSeek is moving toward the public markets. The timeline will put it alongside OpenAI. The firm recently pushed back its IPO to 2027 and valued it at $1 trillion.
Why investors keep paying
DeepSeek isn’t just repricing for hype. Its annual revenue recently reached $400-500 million from cloud access to its models, The Information reports. In June, the lab accounted for about 23% of enterprise AI gateway tokens processed by Vercel. According to TechCrunch. Only Anthropic, at 32%, ranks higher.
He built this position on open source reasoning models. Although they perform far from US border laboratories Export controls restricting China’s access to advanced Nvidia chips. Instead, DeepSeek’s cloud service runs on Huawei’s hardware. It’s one of the clearest signs that a competitive AI stack can function without American silicon.
A founder who answers to no one
The evaluation has been personally transformative for Liang. Its stake, which is still about 78%, is worth $36 billion, up from $16.7 billion on paper. This figure comes from the Bloomberg Billionaires Index. He does it the world’s richest artificial intelligence developerAhead of Anthropic’s Dario Amodei and OpenAI’s Greg Brockman.
Liang told potential investors that DeepSeek would prioritize long-term research over rapid commercialization. It will continue to release open source models down the road artificial general intelligence. It’s easier to promise when your supporters can’t vote more than you. Whether or not DeepSeek can respond to public shareholders once is the real question that the DeepSeek IPO will test.
