Hangzhou‑based AI lab DeepSeek announced its first external fundraising effort, aiming to raise a minimum of $300 million while positioning itself for a 70 billion‑yuan (~$10 billion) valuation. Founder Liang Wenfeng used the pitch to underscore a long‑term ambition: building artificial general intelligence (AGI) and keeping its models open source, rather than chasing short‑term commercial deals.
Until now, DeepSeek operated on capital supplied by High‑Flyer Quant, the quantitative trading firm Liang founded and that continues to back the lab. Liang has previously described the lack of outside investors as a deliberate shield against product‑roadmap pressure. The new round marks a departure, driven by the sheer compute cost of training models at DeepSeek’s scale – expenses that even a profitable hedge fund cannot shoulder alone.
DeepSeek’s recent releases illustrate its research‑first ethos. In April, the lab unveiled the V4‑Pro and V4‑Flash models, a 1.6‑trillion‑parameter mixture‑of‑experts system and a 284‑billion‑parameter variant, respectively. Both were released under permissive open‑source licenses and optimized for Huawei Ascend and Cambricon silicon as well as Nvidia GPUs, signaling a clear tilt toward the domestic Chinese market, which faces growing restrictions on U.S. accelerator technology.
The company’s pitch to investors leans heavily on the success of its R1 model, launched in January 2025. R1’s debut reportedly erased roughly $600 billion from Nvidia’s market capitalization in a single trading session, a claim that sparked debate but underscored the strategic point: a Chinese lab could compete at the frontier of AI while keeping costs dramatically lower than U.S. counterparts.
Details of the financing terms remain private. Investor identities have not been disclosed, and DeepSeek has not provided on‑record comment for Bloomberg’s coverage. The valuation figure represents the target price DeepSeek hopes to achieve, not the amount of capital it expects to secure. The round, when closed, will be the first time external capital aligns with the lab’s stated AGI‑first, open‑source agenda.
Regulatory scrutiny adds another layer of complexity. Chinese authorities have spent the past two years defining oversight mechanisms for foundation‑model developers. A financing deal of this magnitude, coupled with an explicit AGI ambition, is likely to attract attention from regulators monitoring the rapid expansion of domestic AI capabilities.
Liang’s public commitment signals a clear direction for DeepSeek: continue releasing cutting‑edge models under open licenses, prioritize research over immediate profit, and aim for the lofty goal of AGI. Whether the market will back that vision with the projected $300 million—and whether the lab can sustain its open‑source stance amid mounting competitive and regulatory pressures—remains to be seen.
This article was written with the assistance of AI.
News Factory SEO helps you automate news content for your site.