DeepSeek founder Liang Wenfeng has told prospective investors that the Hangzhou AI lab is targeting AGI as its primary goal, in the company's first-ever external funding round. The raise targets a $10bn valuation, with at least $300m sought in this round and potentially $7bn or more long-term. Previously self-funded through High-Flyer Quant, DeepSeek is opening to outside capital as training costs exceed what even a profitable hedge fund can cover. Liang has committed to keeping models open-source under permissive licensing and prioritizing frontier research over short-term commercialization. The lab's recent V4-Pro and V4-Flash models, a 1.6-trillion-parameter MoE system, were released open-source and optimized for both Nvidia and Chinese domestic chips like Huawei Ascend.
