Chinese artificial intelligence startup DeepSeek is preparing for an initial public offering, with a potential filing as early as late 2026 and a market debut targeted for 2027, according to multiple reports. The Hangzhou-based company is currently working with accounting and banking advisors on a possible mainland China listing, while simultaneously pursuing additional private funding.
The planned IPO comes amid a sharp rise in DeepSeek’s valuation and revenue, underscoring intensifying competition among AI firms to secure capital for compute infrastructure, data centers, and advanced chips.
Valuation climbs toward $70billion
DeepSeek recently completed a $7billion funding round and is now in discussions to raise fresh capital at a valuation between $71billion and $74billion. The rapid succession of fundraising efforts reflects both strong investor demand and the capital-intensive nature of large-scale AI model development.
The company’s annualized revenue has climbed to an estimated $400million to $500million, driven largely by cloud-based access to its AI models. This positions DeepSeek among a growing cohort of AI firms generating meaningful commercial income, rather than relying purely on research funding.
While the final listing venue has not been confirmed, the company is reportedly considering mainland exchanges such as Shanghai’s STAR Market, which is designed to support high-growth technology firms.
Founder Liang Wenfeng emerges as richest AI builder
DeepSeek’s rise has also reshaped the global AI wealth landscape. Founder Liang Wenfeng has become the world’s richest creator of AI models, with a net worth estimated at $36billion following the company’s latest funding round.
Liang’s wealth surpasses that of prominent Western AI figures, including OpenAI president Greg Brockman and Anthropic co-founder Dario Amodei. Notably, OpenAI CEO Sam Altman is absent from comparable rankings due to holding no direct equity in OpenAI.
A key factor behind Liang’s outsized wealth is his unusually high ownership stake. He is estimated to control between 78% and 84% of DeepSeek, a level of founder equity rarely seen in Silicon Valley, where venture capital funding typically leads to significant dilution.
A different capital model
DeepSeek’s ownership structure reflects its unconventional origins. Liang initially funded the company using profits from Zhejiang High-Flyer Asset Management, a quantitative hedge fund he founded. This allowed DeepSeek to scale early without relying heavily on external venture capital.
The result is a markedly different capital model from that of many US-based AI firms, where multiple funding rounds often reduce founder control. DeepSeek’s approach has enabled Liang to retain both strategic authority and a dominant financial stake as the company scales.
Race for AI Infrastructure
The company’s IPO preparations highlight a broader trend across the AI sector: an accelerating race to secure funding for infrastructure. Training and deploying advanced models requires vast computational resources, making access to capital a critical competitive factor.
DeepSeek’s dual-track strategy — raising private capital while preparing for a public listing — suggests it is positioning itself to compete at the highest level of global AI development, where scale, speed, and funding capacity increasingly define market leaders.
