DeepSeek Founder Declares AGI as Goal as $10bn Funding Round Advances
May 22, 2026 – 7:56 am
Liang Wenfeng tells prospective investors the Hangzhou lab will prioritize frontier research over revenue, in its first-ever external raise.
DeepSeek founder Liang Wenfeng has informed potential investors that his company aims to pursue artificial general intelligence (AGI) as a primary objective and will continue releasing open-source models rather than focusing on short-term commercialization. This message was conveyed during the ongoing $10 billion funding round, as reported by Bloomberg.
The startup is seeking at least $300 million in external capital this round, according to The Information. The long-term target could reach $7 billion or more as DeepSeek advances toward recurring revenue. This is the first time Liang has accepted outside investment.
This shift is primarily due to scale; training models of DeepSeek’s size is beyond what even a profitable hedge fund can finance independently. The AGI framing also signifies where DeepSeek positions itself in the market.
In April, DeepSeek released its V4-Pro and V4-Flash models—a 1.6 trillion parameter Mixture-of-Experts system and a smaller 284 billion parameter variant, both open-sourced under permissive licenses. The V4 family is designed to run on various hardware, including Huawei Ascend, Cambrion, and Nvidia, targeting the domestic Chinese market.
The pitch to investors emphasizes the research-first identity that led to the groundbreaking R1 model in January 2025, which caused a significant drop in Nvidia’s market capitalization. The strategic impact was clear: a Chinese lab could compete at the frontier level while keeping costs low through open-source practices.
While specific terms are yet to be publicly disclosed (including investor identities, official valuation, and close date), DeepSeek has confirmed its direction of travel: open-source, frontier-oriented, and AGI as its primary focus rather than enterprise revenue.