DeepSeek, the Hangzhou-based open-weight AI lab, resumed fundraising in early August at a $74 billion pre-money valuation target after pausing a similar round in July. The company is seeking up to $8 billion in fresh capital with Monolith Management in talks to participate. Founder and CEO Liang Wenfeng is committing up to 20 billion yuan (~$3 billion) personally, which would lift his direct equity stake from 1% to 34% and total control (including indirect holdings) to approximately 84.29%.
DeepSeek closed its maiden external fundraising round at $50 billion valuation in June 2026, raising approximately $7.4 billion from investors including Tencent, CATL, and the state-backed National Integrated Circuit Industry Investment Fund. The company was founded in July 2023 and operated on capital from Liang's quantitative hedge fund High-Flyer until seeking external investment. A July regulatory filing by Anhui Korrun (a Chinese luggage maker) revealed an indirect stake that implied a post-money valuation above $52 billion.
DeepSeek has begun preparations for a Shanghai STAR Market listing targeting 2027, with reports indicating the company may file listing documents as soon as end-2026. The company is not pursuing traditional VC growth financing but rather appears to be building out in-house infrastructure capital at scale. Proceeds from the current round are earmarked for data center buildouts.
For practitioners: DeepSeek's $50B+ valuation is unusual because it is not pricing revenue, customer base, or installed base—the company has essentially zero commercial revenue. It is instead pricing infrastructure arbitrage: the founder's track record (High-Flyer is a successful quant fund), DeepSeek's demonstrated cost efficiency in model training (V3 and R1 models published with training costs far below Western equivalents), and capital durability. The structure (founder control via limited partnership) mimics High-Flyer's ownership, suggesting long-term commitment to R&D over quick exits.