Chinese AI star DeepSeek eyes $10 billion valuation in first external funding round to set option pricing and retain core talent

While DeepSeek and its parent company, High-Flyer Capital Management (幻方量化), have not officially commented, sources from state-owned equity institutions indicate the news is highly credible, noting that investment slots are extremely scarce and largely inaccessible to ordinary institutions.

This move marks a major strategic shift for DeepSeek, which has long adhered to a "no external capital" policy. Industry analysts suggest that, beyond surging computing costs and the massive capital required for top-tier model R&D, talent attrition is a primary driver behind the move. Recently, several core researchers have departed, including Luo Fuli (罗福莉), a key contributor to the V2 architecture who joined Xiaomi (小米), and Guo Daya (郭达雅), a top researcher who joined ByteDance’s (字节跳动) Seed team. Investors believe a key goal of this round is to establish a clear market pricing anchor for employee options, stabilizing the team by increasing the liquidity and premium appeal of their equity.

On the technical front, DeepSeek maintains an aggressive R&D pace. Reports indicate that DeepSeek is expected to officially release its next-generation trillion-parameter model, DeepSeek-V4, by late April. While some worry that external capital might affect the founding team’s independence, securing capital reserves and talent incentives has become essential as the Chinese AI industry enters a deep commercialization phase. The financing terms are reportedly being personally overseen by founder Liang Wenfeng (梁文锋) under extremely rigorous criteria, with domestic capital expected to lead the round.