In the heart of Silicon Valley, a phenomenon is reshaping the artificial intelligence job market: the obsessive search for a handful of brilliant minds capable of pushing technology to superintelligence levels. According to a report by The Wall Street Journal, Mark Zuckerberg has spent months compiling what is internally known as “The List,” a secret catalog of elite researchers and scientists, to whom he is offering compensation packages that can reach as high as $100 million.

Meta’s bet — the parent company of Facebook and Instagram — is no small move. After its latest generative AI model, Llama 4, received a lukewarm reception, Zuckerberg decided to double down on his efforts to recruit those he sees as the architects of the digital future. The goal: to create a lab capable of developing systems smarter than humans.

An exclusive and multimillion-dollar club

The profile of these candidates is as exclusive as it is costly: Ph.D.s in mathematics or computer science from universities like Berkeley, Stanford, or Carnegie Mellon, with experience at companies such as OpenAI, Google DeepMind, and Anthropic. They often know each other, share research, and compare offers as they decide whether to move to Meta or stay at their current firms. Some even negotiate their recruitment as a “package,” bringing trusted colleagues along.

The American newspaper recounts the case of Lucas Beyer, a European researcher who worked at Google and OpenAI before receiving a direct offer from Zuckerberg. Although not all recruits reach nine-figure sums, the amounts are still astronomical, especially compared to typical tech industry salaries.

Zuckerberg himself leads the conversations with candidates. According to the investigation, he keeps a chat titled “Recruiting Party” with two Meta executives, where they exchange strategies and discuss the best way to approach each prospect: email, text message, or WhatsApp.

A race with the flavor of espionage

In this environment of fierce competition, companies have tightened precautions to protect their intellectual property. OpenAI and Anthropic operate on access-restricted floors with blinds always drawn, while at Safe Superintelligence, interviewees must leave their phones in a Faraday cage that blocks signals.

The tension even reaches national security: The Wall Street Journal reports that Anthropic invited an FBI agent to warn staff about the risk of foreign espionage.

At the same time, top executives like Sam Altman (OpenAI), Sundar Pichai (Google), and Elon Musk (xAI) are personally courting the same talent. Behind this attention lies a simple reason: the most experienced researchers accumulate tacit knowledge — what some call “tribal knowledge” — that is nearly impossible to replicate, a strategic asset that can make the difference in the race for advanced artificial intelligence.

The power of infrastructure

For many of these scientists, money is just one more incentive. What truly attracts them are the nearly unlimited computing resources that only a few big tech companies can provide. Meta’s planned investment in AI this year amounts to around $70 billion. Compared to that figure, the experts seem like a bargain: a handful of star engineers cost much less than a few data centers.

According to The Wall Street Journal, many of the recruits until recently preferred to pursue academic careers as university professors. Today, the chance to lead large-scale experiments and help shape the evolution of artificial intelligence is simply too attractive to remain in academia.

While Zuckerberg keeps adding names to his list and organizing private meetings in Palo Alto and Lake Tahoe, the question remains whether this salary war will ultimately concentrate even more technological power in the hands of a small group of companies that can afford everything.