Why AI superathletes could be winning $100 million bonuses in Silicon Valley

SAN FRANCISCO — Meta CEO Mark Zuckerberg has been on a recruiting spree in recent weeks, luring top artificial-intelligence researchers and engineers away from OpenAI, Apple, and Google.

OpenAI CEO Sam Altman said in a podcast interview that the “crazy” campaign involved offering his staff $100 million signing bonuses. Tech podcast TBPN marks the latest industry moves by posting baseball card-style graphics featuring AI superstars on social media with the text “TRADED.”

For those outside of Silicon Valley, it raises the question: What do AI superathletes know that makes these relatively obscure workers so highly prized?

A $100 million bonus would dwarf the $17.1 million median pay of S&P CEOs in 2024, according to Equilar and the Associated Press. Meta’s recent recruits include an engineer who previously led Apple’s AI models team, whose multiyear contract is worth $200 million, according to Bloomberg News.

Apple CEO Tim Cook, the most highly paid among leaders of the largest tech firms, received $63 million in 2024. Apple and OpenAI did not respond to requests for comment.

Meta’s hiring blitz began in June as it launched a new team dedicated to creating superintelligence, a term for machines hoped to one day outperform humans at every possible task. It is led by Alexandr Wang, who recently became the company’s chief AI officer after it agreed to acquire his start-up Scale AI for almost $15 billion.

“For our superintelligence effort, I’m focused on building the most elite and talent-dense team in the industry,” Zuckerberg wrote in a social post Monday.

He responded to recent reports of his compensation offers in an interview posted by The Information on YouTube on Tuesday, saying that “a lot of the numbers specifically have been inaccurate” but acknowledging there is “an absolute premium for the best and most talented people.”

A list of recent recruits posted by Wang on X last month provides a snapshot of the skills and experience of AI superstars. Of the 11 people on the list, seven previously worked at OpenAI. (The Washington Post has a content partnership with OpenAI.)

The majority of Wang’s list were men, mirroring the entrenched gender imbalance in the tech industry, which some evidence has suggested is more pronounced in AI. And most people named had completed undergraduate studies outside the United States, according to their LinkedIn pages, reflecting how top-tier AI labs pull in talent from around the globe.

Silicon Valley companies employ thousands of workers in the U.S. on H-1B visas, a category that allows hiring overseas for jobs requiring specialized knowledge. Data on H-1B applications for positions with the job title “machine learning scientist,” capturing all levels of talent, show that the highest base salaries offered jumped last year.

Seven of Meta’s superstar 11 completed their undergraduate studies at Chinese universities, according to their LinkedIn profiles, before continuing their studies in the United States. Four attended Beijing’s prestigious Tsinghua University.

China’s government and universities have invested heavily in training the next generation of AI talent as part of a national push to overtake the U.S. in the crucial technology. While the U.S. leads in producing the top AI software, China had notched more AI publications and patents, according to Stanford University’s 2025 AI index.

Many of the new hires on Wang’s list earned Ph.D.s from top U.S. universities such as the Massachusetts Institute of Technology, Carnegie Mellon, and Stanford. They’ve completed internships or held jobs at companies with major AI research programs such as Google, OpenAI, or Anthropic, which offers the Claude chatbot. Several people on the list have previously interned or worked at Meta.

Working at those companies can provide experience with some of the largest and most sophisticated computing infrastructure on the planet.

The resumés of Meta’s hires reflect the cocktail of technical mastery needed to become a top name in AI development. To reach that level a person needs to be a strong software engineer as well as adept in statistics, linear algebra, probability, and other math underpinning machine learning algorithms, said Daniela Rus, director of MIT’s Computer Science and Artificial Intelligence Laboratory.

AI superstars have to put it all into practice on giant computer systems that occupy thousands of powerful graphics chips to create new machine learning models. “People have trained for years to have this mastery,” she said. “There really is a limited talent pool that [can] push AI forward.”

Rus added that top AI achievers also bring a dash of artistry to those technical domains. “My students like to be thought of as … creatives,” she said. “There is an art form to building an AI solution.”

To make a breakthrough new AI product like ChatGPT can require combining research ideas that rethink foundations of the field with practical flair, to create a system unlike what came before that can scale up and serve the world.

The value of these rare individuals to tech companies has created a class of recruiters dedicated to headhunting AI mavens. Amy Anton, vice president of AI talent for Lightspeed Venture Partners, said although the contest for hires has been heating up in recent weeks, AI experts have long been highly sought-after by Big Tech companies.

“Research scientists drive the breakthrough innovations that can create entirely new product categories,” Anton said. She previously worked as the head of North American talent acquisition for Google’s DeepMind AI lab, where job descriptions for AI scientists frequently sought individuals who could “create fundamentally new ways of doing things,” she recalled.

The hefty compensation companies like Meta are willing to pay can be explained in part because snagging a big name can attract other talent, Anton said — like a football squad drafting a star quarterback who becomes a magnet for others.

Top names bring credibility and clout to a company, Anton said. But she added that even a highly talented team needs cohesion, not just research power, to come up with market-leading innovations. “It’s not great to just have a trophy case of researchers,” she said.

Executives and CEOs need to create the right conditions for stars to get things done and feel they are doing impactful work. And public contests for high-priced recruits can leave existing staff feeling undervalued, Anton said. “It’s possible that Meta will have attrition problems,” she said.

Zuckerberg may have had those challenges in mind when he emphasized the technical advantages Meta offers all its researchers in his interview with The Information, saying the company provides ample computing resources.

“Having basically the most compute per researcher is a strategic advantage — not just for doing the work, but for attracting the best people,” he said.

Avi Goldfarb, a professor who studies artificial intelligence at University of Toronto’s business school, says for a company such as Meta, which reported $62.4 billion in profit for 2024, $100 million could be reasonable for an AI superstar who delivers the huge advances tech leaders hope AI will bring.

“If having the right person in the right place at the right time can increase those numbers by 10%, then that person is underpaid if they’re making $100 million,” Goldfarb said. But he added that since it’s unknown just how transformational AI will prove to be, “it’s not obvious that they’re being over- or underpaid.”

Goldfarb says sports is an apt analogy for AI development, in that the most talented people can open new opportunities for their employers and teammates. But in sports only one team wins the championship every year. “In business, there can be lots of winners,” he said, and it’s unclear whether AI is “a winner-take-all market.”

Zuckerberg’s recent hires and other comments last week suggest he’s not taking any chances of being left behind. He announced plans for a giant data center campus large enough to obscure Manhattan to power future AI projects by his superintelligence team.