"What is the most expensive in the 21st century? Talent!" - This profound truth, articulated by Ge You's character "Uncle Li" in "A World Without Thieves" two decades ago, continues to resonate today. In the forefront of the AI race that will shape the future technological landscape, Silicon Valley giants are maniacally proving this quote right: top AI talent has emerged as a strategic resource rarer than gold and more pivotal than computing power, sparking an intensified "arms race."

Recently, tech titans such as Meta, Google, Apple, and Musk's xAI have spared no expense in staging one astonishing "war for talent" after another. Meta was exposed for successfully luring four core researchers from OpenAI, all key contributors to the GPT-4 and o series models. Google, on the other hand, shelled out a whopping $2.4 billion to "integrate" the core team of AI startup Windsurf. Even more astonishingly, rumors suggest that Meta offered a staggering $100 million signing bonus in an attempt to poach top talent from OpenAI. This is not merely about hiring at exorbitant salaries; it has evolved into "talent vacuum" looting through astronomical salaries and acquisition of entire teams.

Behind these giants' reckless spending lies a stark revelation: the extreme scarcity of top AI talent. Globally, there are fewer than 1,000 experts who can truly be labeled as "top," and their value is highly esteemed in Silicon Valley. They are not just the brains behind model construction and algorithm optimization but also directly determine the upper limits of large model capabilities, serving as indispensable "game changers" for giants on the path to Artificial General Intelligence (AGI). In this race to dominate the future, whoever secures top talent holds the first-mover advantage in building technical barriers, driving data flywheels, and defining industry standards. Losing them means losing the future.

01. Silicon Valley Giants Scramble for Top AI Talent

According to US tech media Wired, four OpenAI researchers will join Meta: Shengjia Zhao, Shuchao Bi, Jiahui Yu, and Hongyu Ren. Zhao Shengjia is the core developer of OpenAI's reasoning models o1-mini and o3-mini; Yu Jiahui leads OpenAI's perception team, overseeing projects like o3, o4-mini, GPT-4.1, and GPT-4o; Bi Shuchao heads OpenAI's multimodal post-training team; and Ren Hongyu leads post-training for the o3 and o4-mini models and is responsible for the open-source model project to be released this summer.

This move is the latest in a series of aggressive initiatives by Meta CEO Mark Zuckerberg, who aims to catch up with OpenAI, Anthropic, and Google in building Artificial General Intelligence (AGI).

In Meta's talent poaching spree, exorbitant salaries have been frequently mentioned. Notably, in an interview in early June, OpenAI CEO Sam Altman revealed that Meta had tried to poach OpenAI employees with signing bonuses of up to $100 million and higher annual salary packages, but "so far, our best talent has not decided to accept their invitation."

Beyond astronomical salaries, Silicon Valley giants also use mergers and acquisitions to amass top AI talent. On July 11, it was revealed that Google DeepMind had successfully integrated the core team of AI startup Windsurf. Google will pay $2.4 billion (approximately RMB 17 billion) in license fees and compensation for Windsurf co-founder Douglas Chen and some senior researchers to join Google and contribute to its AI programming projects.

Similarly, Meta recently acquired nearly half of Scale AI's shares at a huge cost and brought in its young CEO to serve as its Chief AI Officer.

Whether it's Meta, Google, Apple, or Musk's xAI, they are all poaching talent, either by acquiring entire star startup teams or directly poaching from OpenAI and Anthropic.

02. The Scarcity of Top AI Talent

The lengths to which giants are going to poach talent deeply reflect the scarcity of top AI talent. According to a report by Forbes, since 2019, the number of job postings for AI skills has grown by 21% annually globally, far outpacing the supply of AI talent. Fortune magazine further points out that "global top AI experts are scarce resources, with currently fewer than 1,000 in total worldwide." Bain & Company predicts that the AI talent gap is expected to persist until 2027 and will be difficult to fill in the short term.

Fortune magazine's assertion that there are fewer than 1,000 top AI experts globally underscores the extreme scarcity of talent in the AI field. Why is Silicon Valley so lacking in top AI talent? Industry insiders explain that "there are actually many AI talents, but labeling in Silicon Valley is very strict. Only those who have worked in the core teams of companies like OpenAI, Google, and Meta and served as team leaders can be called 'top AI talents.' Otherwise, even Ph.D.s from Stanford and Harvard may not secure such high salaries."

Moreover, under the premise that Silicon Valley giants generally use NVIDIA computing cards, top AI talents actually determine the upper limits of large models.

For instance, in data screening and processing, AI can handle large amounts of data but cannot judge its quality and relevance. Human judgment becomes crucial here. High-quality data input ensures that AI output has reference value. Data screening, cleaning, and processing all rely on human experience and judgment.

In model selection and optimization, the core of AI lies in model construction and optimization. Different tasks require different models, and model effectiveness depends on continuous optimization. Human judgment plays a decisive role in this process. Which model is more suitable for the current task? How to adjust model parameters for optimal results? These questions demand professional judgment and rich experience.

03. No Giant Wants to Lose the First-Mover Advantage

From a business perspective, Silicon Valley giants hoard top AI talent because none wants to lose the first-mover advantage.

R&D investment in the AI field (especially large models) is substantial, and technological breakthroughs are highly complex and uncertain. Achieving breakthroughs first (such as more powerful model architectures, higher training efficiency, and better inference performance) can establish significant technical barriers, pulling ahead of competitors.

Specifically, leading AI products and services can attract more users, generating high-quality user interaction data. These data are crucial for training and iteratively improving models. Early movers can initiate and accelerate this "data flywheel" earlier, forming an insurmountable data advantage.

Moreover, establishing platforms first (such as APIs, developer tools, and app stores) can attract developers and enterprise users to build applications and services on top, fostering a thriving ecosystem and increasing user stickiness and switching costs. Market leaders can define user expectations, shape industry standards, and establish strong brand recognition, which has a huge advantage in attracting customers, partners, and subsequent talent.

To achieve these goals, top AI talent is indispensable. Competition in the AI field is essentially a race for innovation speed. Top talents (top researchers, engineers, architects) are the core sources of groundbreaking ideas and efficient engineering implementations. They solve complex problems faster, push model capabilities, and are the key to maintaining technological leadership. By locking in these talents through high salaries, superior research environments, massive computing resources, and influence, not only can one strengthen oneself but also effectively weaken competitors, delaying or even preventing their pursuit. This is an active "defensive" strategy.

Therefore, in this crucial race to define the future of AI, the first-mover advantage translates into technological dominance, market share, and huge commercial value. Top AI talents are the ultimate "weapons" and "fuel" for creating, maintaining, and expanding this advantage. The giants deeply understand that losing talent equals losing the spark of innovation and the motivation to move forward, ultimately leading to the loss of the first-mover advantage. Hence, the relentless competition and hoarding of top talent are not accidental but a strategic necessity to maintain their survival and dominant position amidst fierce competition. The war for talent is the core battlefield where giants defend their future.

- End -