It is reported that DeepSeek made a "no poaching" requirement to investors: adding a talent lock in addition to the triple lock

According to CNBC, DeepSeek founder Liang Wenfeng made a special request to potential investors in an online investor meeting in May this year: One of the prerequisites for investing in DeepSeek is a promise not to poach DeepSeek’s employees or encourage them to start a new business. This news has not yet been officially confirmed and should be treated with caution. But if true, this will form a complete closed loop with the triple control insurance previously designed by Liang Wenfeng of "no voting rights, no board seats, and a five-year lock-in period" - plus an additional "talent lock."

DeepSeek has rejected external financing since its inception, hoping to focus on research rather than commercial operations. But after losing some of its core talents, the company finally decided to launch a financing plan. The list of brain drain is worrying: V3 core developer Luo Fuli left at the end of last year and was poached by Lei Jun to Xiaomi with an annual salary of tens of millions. The model subsequently launched by her MiMo AI team has surpassed DeepSeek similar products in benchmark tests; R1 core researcher Guo Daya joined Byte; Wang Bingxuan, the core author of the first generation large language model, went to Tencent. In addition, Tencent also hired Yao Shunyu, who had worked at OpenAI, as its chief AI scientist last year. In order to obtain stronger AI R&D capabilities, Chinese companies are attracting overseas experienced talents in various ways.

The loss of these names is not only the departure of individual talents, but also reveals the structural dilemma faced by DeepSeek: there is no circulation market for options in unlisted companies, and the incentives in the hands of employees are equivalent to paper that cannot be cashed. One of the core purposes of this round of financing - to market-based pricing for employee options - and the "no poaching" clause point to the same problem: at a time when the competition for AI talents is fierce, the core competitive element of Chinese AI companies is not computing power or data, but people who can create the next generation model.