The artificial intelligence boom is reshaping the boundaries of compensation, ambition, and scientific prestige. Meta’s recent offer of a $250 million package to AI researcher Matt Deitke—reportedly with $100 million in the first year alone—marks a turning point in Silicon Valley’s escalating talent war, surpassing not only the pay of today’s tech professionals and athletes, but even the rewards once granted for humanity’s greatest scientific achievements.
Deitke, a 24-year-old AI researcher who previously led the multimodal system Molmo at the Allen Institute for AI and co-founded the startup Vercept, specializes in systems that process images, sound, and text together—exactly the kind of next-generation AI Meta is racing to develop. His expertise made him a prime target. And he’s not alone. Meta CEO Mark Zuckerberg reportedly offered another top engineer a staggering $1 billion to lure them into the company’s AGI efforts.
Behind these astronomical figures is a race for artificial general intelligence—AGI—or what some in Silicon Valley call superintelligence: machines capable of performing intellectual tasks at or beyond the level of humans. Tech giants like Meta, OpenAI, Google DeepMind, and Anthropic believe that the company that gets there first will hold the keys to the future, unlocking not only dominance in AI but the ability to reinvent products, create entire industries, and automate vast swaths of the global knowledge economy.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
Zuckerberg recently told investors that Meta would continue pouring resources into AI research “because we have conviction that superintelligence is going to improve every aspect of what we do.” In an open letter, he called it “an exciting new era of individual empowerment,” though he stopped short of defining exactly what superintelligence means.
But the size of Meta’s investment makes its priorities clear. The company plans to spend over $80 billion on capital expenditures this year, mostly aimed at AI. One senior executive told The New York Times, “If I’m Zuck…is it worth kicking in another $5 billion or more to acquire a truly world-class team to bring the company to the next level? The answer is obviously yes.”
A New Standard of Scientific Compensation
According to Ars Technica, the numbers leave even history’s most iconic scientists in the dust. J. Robert Oppenheimer, who led the Manhattan Project that developed the atomic bomb, made about $10,000 a year in 1943—about $191,000 today. Deitke’s offer is over 300 times that. Even Thomas Watson Sr., IBM’s CEO in 1941 and one of the wealthiest executives of the time, received what would be $11.8 million in today’s money, less than a quarter of Deitke’s annualized package.
During the Apollo program, Neil Armstrong earned the equivalent of about $245,000 for his historic 1969 moon landing. Today, a top AI researcher at Meta makes that amount in three days.
Historically, even revolutionary scientists like Claude Shannon—the father of information theory—worked on modest salaries at Bell Labs during its golden era. In that era, the pay difference between the director and the lowest-paid technician was about 12 to 1. In today’s AI world, that ratio would be laughable.
Why AI Talent Is So Expensive
Several forces are driving this runaway market. Unlike government-backed mega-projects like Apollo or the Manhattan Project, the AI race is driven by competing trillion-dollar corporations. And the talent pool is tiny. Only a few dozen individuals in the world have deep experience developing frontier multimodal AI systems. Companies are bidding aggressively, sometimes offering tens of thousands of GPUs—specialized hardware required to train massive models—on top of money and equity.
Deitke’s peers, many of whom are still in their 20s, now share offer letters in private Discord and Slack groups and sometimes hire informal agents to negotiate deals. One top researcher was told by recruiters they’d be given access to 30,000 GPUs—a resource that would have taken a national lab to muster just a few years ago.
The belief among executives is that AGI will not just be a better product. It could invent other products, write software, discover scientific breakthroughs, and fundamentally transform entire economies. The stakes are so high that spending hundreds of millions on individual researchers is seen as a justifiable bet. In this context, Deitke’s $250 million offer—or even a rumored $1 billion—may seem extreme, but not irrational.
A Modern Gilded Age
These developments also mark a return to levels of industrial wealth concentration last seen in the Gilded Age. But unlike the steel barons or railroad tycoons of that era, today’s AI firms are valued in the trillions and operate at a scale that affects the entire globe. And while AI promises productivity and transformation, the economic upside is not being evenly distributed.
Meta, for example, has laid off thousands of workers even as it aggressively invests in AI. It’s a pattern seen across the tech industry: aggressive hiring for a tiny cadre of AI researchers and engineers, and widespread job cuts elsewhere. The economic model seems clear—automate what can be automated, reduce costs, and double down on scalable intellectual capital.
More Than Just Money
For some, these record-setting offers raise ethical questions about the priorities of today’s tech industry. Is the AI boom really about human empowerment, as Zuckerberg suggests, or about cornering the next trillion-dollar platform and the power that comes with it?
Whether AGI ever truly materializes remains an open question. But the belief that it might is already reshaping compensation, research priorities, and the nature of scientific work itself. For now, researchers like Matt Deitke represent the sharpest edge of that shift, where science, capital, and ambition collide.
And if he chooses to cash in and walk away in a few years, as some of his peers may, few would blame him. As one of his Vercept co-founders joked after the Meta deal became public, “We look forward to joining Matt on his private island next year.”