China’s AI talent market produced its most talked-about story of the week when Chinese media outlet LatePost reported on April 17 that Guo Daya, a lead researcher on DeepSeek’s R1 model, had joined ByteDance’s Seed AI development team with an annual compensation package approaching 100 million yuan (approximately US$14.7 million). The report spread rapidly across Chinese tech media and social platforms, becoming a proxy for a debate over the true cost of top AI talent in China.
ByteDance moved quickly to contain the narrative. Li Liang, a vice-president of ByteDance’s Douyin Group, posted a public denial on the same day, stating that the Seed team operates under a unified compensation framework that combines a cash salary, ByteDance equity, and Doubao-related stock options. He noted that it is “likely some staff could earn hundreds of millions of yuan by exercising their stock options after four years”, a carefully worded statement that neither confirmed nor denied Guo’s hiring, and did not rule out the possibility of very large total compensation over time. Guo’s name had not appeared in ByteDance’s internal staff directory as of the time of reporting, though a ByteDance employee noted that new staff sometimes use pseudonyms. The South China Morning Post could not independently verify the original LatePost report.
Tencent’s Parallel Poaching Campaign
While the ByteDance-DeepSeek story dominated headlines, a separate, arguably more structurally significant development was unfolding: Tencent has reportedly been aggressively poaching AI researchers from ByteDance, offering to double current salaries and providing substantial premiums for fresh Ph.D. graduates. This bidding war between two of China’s largest tech companies reflects the acute shortage of researchers with the specific combination of skills, large-scale model training, reinforcement learning, and systems engineering, required to compete at the frontier.
The talent competition is not purely domestic. Chinese tech giants are also targeting researchers at US companies, with some Silicon Valley-based Chinese nationals returning to China amid AI layoffs in the US that have disrupted immigration status for those on work visas. A human resources manager at a Silicon Valley startup who previously worked at Baidu and TikTok described the situation: “The two markets are fighting over the same bunch of people.” For Chinese nationals in the US, sudden layoffs also threaten their immigration status, making a return to China’s booming AI sector an increasingly attractive option, particularly when compensation packages offered rival or exceed those of US companies. In addition, EastFrontier has reported that Beijing is working to lure top-tier AI talent back from the US.
What the Talent War Reveals About China’s AI Moment
The intensity of the current hiring competition is a direct consequence of the DeepSeek moment. When DeepSeek’s R1 model demonstrated that a relatively small team with constrained compute could produce frontier-level results, it validated the thesis that individual researcher quality is a decisive variable — not just raw compute or data scale. That realization has made top researchers extraordinarily valuable, and companies are acting accordingly.
For smaller AI startups and research labs, the escalating compensation expectations at the top of the market create a genuine structural challenge. The concentration of talent at ByteDance, Tencent, Alibaba, and a handful of well-funded startups may accelerate the pace of model development at those companies while leaving the broader ecosystem increasingly talent-constrained. The irony is that DeepSeek itself — the company whose research triggered this talent war — remains a relatively lean operation backed by quantitative hedge fund High-Flyer, operating with a fraction of the headcount of its larger rivals. Whether that model of concentrated, high-quality research teams can survive the poaching pressure it has inadvertently created is one of the more interesting questions in China’s AI industry right now. The talent war also raises a broader question about sustainability: at some point, the salary escalation must either plateau or be absorbed into the companies’ cost structures. For now, the bidding continues.
What This Means for China’s Broader AI Ecosystem
The concentration of top AI talent in a small number of well-capitalized companies has implications that extend beyond those companies. Chinese universities have dramatically expanded AI programs in recent years, and the pipeline of trained researchers is growing. But the researchers capable of working at the frontier, those who can design and train the next generation of large language models, remain a small and intensely competed-over group. The current salary escalation reflects a genuine scarcity, not just competitive signaling.
In China’s national AI strategy, the concentration of talent at ByteDance, Tencent, and Alibaba is both a blessing and a curse. These companies have the resources to attract and retain the researchers needed to compete at the global frontier, which serves China’s broader goal of AI leadership. But the concentration also means that smaller companies, universities, and research institutes are increasingly unable to compete for the same talent, potentially slowing the diversification of China’s AI research base. The government has been aware of this dynamic and has invested in national AI research centers and talent programs designed to develop researchers outside the commercial sector. Whether those programs can keep pace with private-sector salary levels is an open question that will shape the structure of China’s AI industry for years to come.