The artificial intelligence industry is currently embroiled in a fierce “talent war.”
Interestingly, top U.S. AI companies are particularly keen on recruiting Chinese talent in their efforts to poach from one another. In fact, Chinese AI experts have long been in high demand in Silicon Valley.
At NVIDIA, Jensen Huang personally recruited two AI experts, Zhu Banghua and Jiao Jiantao, who have made significant contributions in the fields of large-scale model architecture optimization and multi-modal AI chip design. Both of these leading AI experts graduated from Tsinghua University and went on to pursue further studies and work in the United States.
Looking at Meta (formerly Facebook), Mark Zuckerberg offered the highest salaries in the industry to attract talent. According to reports, in June 2025, Meta poached several core researchers from OpenAI, with each receiving a total salary of up to $100 million in their first year. Notably, over half of these teams are Chinese, including Wang Xuezhi (transliterated, all names mentioned hereafter are transliterated), who was involved in the design of the GPT-4 training framework, and Xiao Taide, who specializes in reinforcement learning algorithms.
Additionally, Meta has poached Ruoming Pang, the head of Apple’s AI foundational model team, who previously led the upgrade of Apple’s Siri semantic understanding engine. Meta offered Pang a total compensation package exceeding $200 million, easily surpassing the salary of his former employer, Apple CEO Tim Cook.
Meanwhile, companies like OpenAI, Google, and Microsoft have not slowed down their hiring efforts, also showing a strong preference for Chinese AI experts and even raising industry salaries to attract top talent.
The resumes of these Chinese AI experts share striking similarities: they typically earned their undergraduate degrees from top-tier Chinese universities such as Tsinghua University, Peking University, and Shanghai Jiao Tong University, then pursued master’s and doctoral degrees in the United States, and remained in the U.S. after graduation.
NVIDIA’s Jensen Huang has been particularly effusive in his praise of Chinese AI experts. He has publicly stated that Chinese AI researchers and scientists are among the best in the world. If you visit the offices of Anthropic, OpenAI, or DeepMind, you will find many top-tier talents from China.
It is no exaggeration to say that Chinese individuals have become the most important participants in the entire AI field.
The importance of talent is fully evident at this time. They are not only the key to the success or failure of a company’s future AI competition but also a sensitive factor in capital market price fluctuations and have become the core focus of AI competition between nations.
Why has this situation arisen? The core reason lies in the fact that AI demands extremely high mathematical proficiency, and Chinese individuals’ strength in mathematics is truly remarkable.
In the 2024 International Mathematical Olympiad, while the U.S. team won the championship, the Chinese team also performed exceptionally well, ranking second. Notably, among the six members of the U.S. team, four were Chinese-born individuals. Additionally, in the Physics, Chemistry, and Biology Olympiads, at least half of the U.S. team members were also Chinese.
This fully demonstrates the strong capabilities of Chinese people in mathematics and science and engineering, and these mathematical and scientific knowledge are the key elements needed to break through the bottleneck in the AI industry. Therefore, Chinese experts have become the most important intangible asset in the AI industry.
Jensen Huang of NVIDIA once mentioned: “China is doing an amazing job. Fifty percent of the world’s AI researchers are Chinese. You cannot stop them, and you cannot prevent them from advancing AI development.”
It is foreseeable that, whether now or in the future, Chinese AI experts will remain key participants in global AI development, and may even become decision-makers. Ultimately, this is because only Chinese experts can effortlessly master the complex algorithms in the AI field.
To use an analogy, if artificial intelligence is likened to a kitchen, then data is the ingredients, algorithms are the recipes, computing power is the gas, and AI experts are the chefs. They are responsible for selecting data (ingredients), designing architectural models (recipes), and allocating computational resources (gas), and ultimately, leveraging their strong mathematical and scientific expertise (cooking experience) to produce “delicious and visually appealing” results. It is clear that mathematical and scientific talent (chefs) is crucial.
Therefore, the “Chinese content” in U.S. AI companies will become a key factor in future competition.
Take NVIDIA as an example: CEO Jensen Huang himself is of Chinese descent and fully understands the diligence and solid mathematical and scientific education background of Chinese individuals. As a result, he recruited Zhu Banghua and Jiao Jiantao, both with Tsinghua University backgrounds, to join NVIDIA at high salaries. Additionally, the dynamic computing power scheduling system developed by Lin Da, a Chinese expert who earned his PhD at MIT, has become one of the core technologies of NVIDIA’s DGX system.
In an interview in May 2025, Huang mentioned that NVIDIA has hired “a large number” of world-class AI researchers from China. This is precisely why NVIDIA has the confidence to maintain its leading position in the industry, with its R&D iteration speed consistently ranking first in the industry.
With ample talent reserves, performance naturally follows suit. NVIDIA’s latest quarterly financial report released at the end of May showed continued growth. Revenue for the first quarter of the 2026 fiscal year reached $44.062 billion, up 69% year-over-year; net income was $18.775 billion, up 26% year-over-year. Of this, data center business revenue reached $39.1 billion, up 73% year-over-year; Blackwell architecture chips accounted for nearly 70% of data center revenue, indicating that NVIDIA has driven the transition of AI chips from the Hopper architecture to the Blackwell architecture.
The reason the Blackwell architecture can be rapidly upgraded and iterated in the short term is that artificial intelligence has fully entered the era of high-performance computing, requiring the processing of more and more complex data. For example, in just the past year, token generation has surged by 50 to 100 times; Microsoft processed over 100 trillion tokens in the last quarter, five times the amount from a year ago. The Blackwell architecture was specifically designed to address this exponential increase in tokens.
As such, the pace of change in the AI industry is truly staggering; even a momentary pause could leave one far behind, and the phrase “rapidly evolving” is no exaggeration.
Given the industry’s rapid pace of change and intense competition, it is no surprise that Chinese individuals, renowned worldwide for their relentless drive, are well-suited to thrive in such an environment.
Meta has clearly recognized this. Despite its current strong performance as one of the “Seven Sisters” of the US stock market, it could easily fall behind, much like Intel did in the past. Some companies have already been warned of the risk of falling behind—on July 9, “Woody Sister” in the US suggested that Apple and Google may have already fallen behind in the AI wave.
Therefore, Meta’s significant investment in Chinese AI experts is aimed at leveraging their strengths to make a strong push in the AI race, and even create a miracle.
In summary, Chinese AI experts have become a key factor in the entire AI competition. While it is certainly cause for celebration that they are thriving in the United States, the AI competition between major powers is a matter of national destiny. It is also hoped that more Chinese AI experts will return to their homeland to contribute to the development of domestic AI.
