1.The War for Talent 

The compensation system for AI talent in China has been turned upside down in barely a year. A few years before, a front-line algorithm researcher saw his annual salary jump from 1 million to 3 million yuan after switching jobs — a leap so dramatic it forced his colleagues to reassess their own market worth. Today, a 3-million-yuan front-line researcher no longer raises eyebrows in the industry. Prices have moved even faster at the base of the pyramid: PhD interns now make the equivalent of 100,000–120,000 yuan per month, whereas just two years ago the average monthly salary for a PhD intern was roughly 10,000 yuan. From the tip of the pyramid to its base, the entire pricing structure for AI talent has been rewritten in a little over a year. 

What companies are really fighting over is not salary but the "formula" — the tacit, unpatentable knowledge of model architecture, data ratios, and training strategy that lives only in the heads of a small number of researchers. Fewer than 200 people in China are capable of leading large-scale pre-training, a complex, high-stakes process of building foundational AI models from massive data and computing power. That scarcity is what makes a company fight over and retain the talent. In late 2025, ByteDance announced that it would increase bonus allocations by 35% and salary adjustments by 1.5 times for AI employees within the company. That same year, Tencent handed its Hunyuan model to 28-year-old Yao Shunyu — within a year he had pulled in eight core researchers, pushed out the old guard, and lifted Hunyuanfrom "lagging" to "top-tier open-source" status. 

2. Tactics Have Escalated Beyond Conventional Business Competition 

Money and authority are only the opening moves. In China, the tactics for poaching talent have multiplied into something far stranger than a conventional bidding war. Some companies don't only compete with salary — they compete with a faculty appointment. Z. ai is the pioneer of this "academic track." Its founder and chief scientist, Tang Jie, along with most of the company's core leadership, comes out of Tsinghua, and the academic atmosphere this creates runs so deep that many Tsinghua students now treat joining Z. ai as the de facto entry ticket to studying under Tang Jie for a master's or PhD. The company isn't hiring graduates so much as pre-recruiting them through the university itself. 

Then there's "cutting off the supply at the root." When a company aims to incubate a rival's model talent, it identifies a competitor's ten most important people and works through them one by one. If they're willing to jump ship, that's plan A. If they're not, plan B is to encourage them to start their own company instead, and help them raise the money to do it. Either outcome weakens the rival. It’s a "negative-round investment": backing a founder before they've even decided to found anything. Some investors approach ByteDance employees while they're still employed, quietly building out a founding team and lining up a first check in advance; others simply station themselves outside Tsinghua and Peking University, cultivating professors, talking to students, and even funding student startups before those students have graduated. 

Poaching talent is only half the job; the other half is making sure it doesn't walk right back out the door. Some AI companies have turned onboarding itself into something closer to an auction. A candidate who accepts an offer and starts within one week receives a three-month signing bonus; within two weeks, that drops to two months; past three weeks, the bonus disappears entirely. This strategy is a countdown clock designed to compress a candidate's hesitation window before a rival, or a moment of doubt, can talk them out of the move. 

3. The Management of Talent 

Once talent has been secured, HR's role shifts from acquisition to management. In China, human resources management is shifting toward digitalization, the creation of a unified digital workplace through next-generation technologies such as mobile internet, cloud computing, big data, and artificial intelligence. The main trends in HR digitization in China are as follows: 

3.1 AI-Driven Automation Becomes the Foundational Infrastructure 

By 2026, AI will serve as the core driver for end-to-end HR business process automation. Highly standardized tasks, such as recruitment, onboarding, transfers, resignations, contract generation, scheduling, administrative services, policy interpretation, and process approvals, will be largely handled by AI and automated systems. As a result, the role of HR will undergo a structural shift, transitioning from "manually executing processes" to "designing processes, defining rules, and overseeing the quality of automation." Automation will become the foundational engineering effort for every company to improve efficiency and reduce operating costs. 

3.2 HR Agents Enter a Phase of Practical Deployment and Expansion 

By 2026, HR agents will become one of the most practical and valuable applications in HR technology. Unlike the robotic process automation (RPA) of the past, HR agents can not only execute tasks but also understand context, generate content, and coordinate tasks across multiple systems. They will gradually take on repetitive and semi-structured tasks such as employee consultations, service ticket processing, drafting performance review materials, driving the recruitment process, and producing learning content. This means the future organizational structure of HR teams will include both "human roles" and "virtual roles." Managers need to rethink how work is assigned, where process boundaries lie, how human review mechanisms operate, and how to collaborate with AI agents. The year 2026 will mark a critical turning point for HR — shifting from "using AI tools" to "working alongside AI." 

3.3 HR Compliance Technology Becomes a Must-Have 

By 2026, digital compliance in human resources will have evolved from an "option" to a "matter of survival." Four government departments, including the Ministry of Human Resources and Social Security, have issued new regulations on platform algorithm negotiations: the 10% cap on labor dispatch is subject to stricter oversight, and standardized reporting of social security contribution bases has been mandated. Currently 

, over 78% of enterprises have listed the selection of HR systems as a core annual budget item, with compliance emerging as the primary decision-making criterion. Nearly 60% of surveyed enterprises explicitly stated that data compliance is the top criterion for eliminating candidates during the system selection process. 

4. Why the Headhunter Matters More, Not Less, in a Digitized Era 

There is a paradox at the heart of China's tech talent market: the more HR management becomes governed by data, the more decisive the human moment of persuasion becomes at the front end. Once a company has built out an AI-driven human resource management system, the acquisition of talent is the one link in the chain that data cannot fully replace — because the decision to leave one's current job is not a rational calculation but an emotional one. A system can automate a contract, but it cannot automate the moment a head-hunter convinces a hesitant researcher that his current company's stock is sinking, his workload is unsustainable, and his talent is being wasted. This is why, even as every other part of the HR function gets absorbed into dashboards and agents, head-hunters have become more central rather than less. This is why, even as our clients embrace HR digitalization, they continue to seek our partnership, as we address the complex human challenges that algorithms are not yet trusted—or equipped—to handle. 

‍