DeepSeek eyes IPO after $71bn valuation talks

DeepSeek is preparing to go public, even as it chases another private raise at a much higher valuation. According to Bloomberg, the Chinese AI lab has begun laying the groundwork for an IPO, with a filing possible as soon as this year and a debut targeted for 2027.

Before that, DeepSeek wants more private capital. The company closed its first-ever outside round only weeks ago, raising about $7bn at a valuation of roughly $50bn. It is now talking to investors about a fresh round at a pre-money valuation of at least 480bn yuan, or about $71bn. Bloomberg says DeepSeek is seeking at least 10bn yuan — around $1.5bn — though the final amount could be several times higher. The Financial Times first reported the fundraising, while The Information said the second round could reach roughly $7.4bn.

DeepSeek valuation rise since April

The speed of the repricing is striking. A $71bn valuation would be up about 40% in six weeks. Since opening to external capital in April at around $10bn, DeepSeek’s valuation has risen fast: talks with Tencent and Alibaba pushed it past $20bn within days, it neared $45bn by May, and the first round closed near $50bn in June.

That first raise was structured unusually. Only one investor — China’s National Artificial Intelligence Industry Investment Fund — received direct equity and voting rights. Other investors put money into a limited partnership controlled by founder Liang Wenfeng, with no voting rights and a five-year lock-up. Bloomberg reported the round was still oversubscribed.

Recommended reading

Hassabis says STEM makes you 10x better at AI

Revenue, chips, and IPO timing

Bloomberg says DeepSeek is working with accounting and banking advisers to finish its financial statements by the end of December, a required step before any filing. A listing could come near the end of this year or in early 2027, depending on when those numbers are ready. No exchange has been confirmed, though mainland China or Hong Kong are seen as the most likely venues. Domestic rivals Zhipu and MiniMax have already listed, and Zhipu’s shares are up nearly 1,600% since its January debut.

Investors appear to be paying for more than momentum. The Information reported that DeepSeek’s annualised revenue recently reached between $400m and $500m, largely from cloud access to its models. In June, the lab accounted for nearly 23% of the enterprise AI gateway tokens processed by Vercel, according to TechCrunch. Only Anthropic, at 32%, ranked higher.

DeepSeek built that position with open-source reasoning models that perform within striking distance of leading US labs, despite export controls limiting China’s access to advanced Nvidia chips. Its cloud service runs on Huawei hardware instead, offering one of the clearest signs yet that a competitive AI stack can be built without American silicon.

That surge has also made Liang far richer on paper. His stake of around 78% is now worth about $36bn, up from $16.7bn, according to the Bloomberg Billionaires Index. Bloomberg says that makes him the world’s wealthiest AI founder, ahead of Anthropic’s Dario Amodei and OpenAI’s Greg Brockman. Liang has told prospective investors that DeepSeek will prioritize long-term research and keep releasing open-source models on the path to artificial general intelligence — a promise made easier by a capital structure that leaves him firmly in control.