China’s artificial intelligence race has entered a new phase, one where access to elite engineers may be as valuable as access to capital.
Chinese AI startup DeepSeek reportedly attached an unusual condition to its first major external fundraising effort: investors must agree not to poach its employees or encourage them to launch competing ventures.
According to a report by fundraising-focused media outlet 36Kr cited by CNBC, founder Liang Wenfeng told prospective investors during a four-hour virtual fundraising meeting in May that any investment in the company would require a commitment not to recruit DeepSeek staff.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
The condition is understood to be attached due to the growing fierce battle for AI talent in China, where technology giants and emerging startups are competing aggressively to secure researchers capable of building advanced artificial intelligence systems and, ultimately, artificial general intelligence (AGI).
The reported demand comes as DeepSeek concluded its first external funding round this week, raising $7.4 billion and securing a valuation exceeding $50 billion. The fundraising makes DeepSeek the most valuable pure-play AI startup in China.
The company’s willingness to seek outside capital marks a significant shift. Since its founding, DeepSeek had largely avoided external funding, preferring to focus on research and model development rather than commercial expansion. However, growing competition for talent appears to have altered that strategy.
The startup has already experienced notable departures.
Among the most significant was the exit of Luo Fuli, a key contributor to DeepSeek’s V3 large language model. Luo left the company late last year to lead MiMo, the AI model team at Xiaomi. Since then, Xiaomi’s AI models have reportedly outperformed some of DeepSeek’s offerings on selected industry benchmarks.
DeepSeek’s concerns are unfolding against a backdrop of intense competition among China’s largest technology companies.
Reports indicate that engineers and researchers are increasingly moving between major AI employers, often with lucrative compensation packages and access to larger computing resources. Earlier this year, ByteDance reportedly lost two prominent AI developers to Tencent. Meanwhile, The Information reported that Tencent invested $20 million in a new AI research laboratory established by Juyang Lin, the former lead researcher behind Alibaba Group’s Qwen models.
Lin announced in March that he was stepping down from the Qwen project, one of China’s leading large language model initiatives.
Alibaba itself has been navigating internal debates over AI strategy. Bloomberg reported in June that the company replaced the head of its enterprise-focused DingTalk platform following disagreements about the unit’s role within Alibaba’s broader AI ambitions.
Chinese firms are also looking beyond domestic talent pools.
Tencent hired Yao Shunyu from OpenAI last year, appointing him chief AI scientist. The move underscored a growing trend among Chinese technology companies to recruit researchers with experience at leading U.S. AI laboratories.
Both Liang and Yao have publicly argued that China should fully commit to pursuing AGI, generally defined as artificial intelligence capable of performing intellectual tasks at or beyond human levels across a wide range of domains.
For DeepSeek, protecting its workforce has become increasingly important as its profile rises globally.
The company burst onto the international AI scene early last year with models that challenged established Western competitors while operating at significantly lower reported training costs. That success elevated DeepSeek from a relatively obscure Chinese research lab into one of the industry’s most closely watched companies.
Its latest fundraising round provides substantial resources to expand computing infrastructure, recruit researchers, and accelerate model development. Yet the reported anti-poaching condition suggests DeepSeek views human capital, rather than funding, as the most critical resource in the next stage of the AI race.
The move also reflects a broader reality across the global AI sector: while billions of dollars continue to flow into model development, the pool of elite researchers capable of building frontier systems remains relatively small.
As competition intensifies between DeepSeek, Tencent, Alibaba, ByteDance, and other AI contenders, retaining those researchers may prove just as important as attracting new investors.