Photo by Diane Serik on Unsplash
Matt Deitke didn’t invent the atomic bomb. He didn’t walk on the Moon. But the deal he just landed dwarfs what those icons earned in their lifetimes—by a lot.
Deitke, a 24-year-old AI researcher who led the development of a multimodal AI system called Molmo and co-founded the startup Vercept, recently accepted a $250 million offer from Meta. That’s right: $250 million over four years.
And that wasn’t even their first bid.
According to reports from The New York Times, Meta CEO Mark Zuckerberg had also dangled a $1 billion offer in front of another unnamed AI engineer. Yep, billion with a “B.” Why? Because in Silicon Valley’s latest tech arms race—building artificial general intelligence (AGI) and so-called “superintelligence”—these engineers aren’t just high earners. They’re the battleground.
Why pay AI researchers more than astronauts and Nobel Prize winners?
Photo by Liz hamburger on Unsplash
The short version: Companies like Meta believe the payoff could be bigger than the internet itself.
Zuckerberg has told investors he’s betting big because he thinks superintelligent AI could “improve every aspect” of Meta’s work. In an open letter, he even said this tech could kick off a new era of “individual empowerment.” That’s a bold-but-vague way of saying: whoever builds machines that think like (or better than) us could reshape the global economy—and maybe even society itself.
It’s not just Meta. Google, OpenAI, and a few others are chasing the same dream. They’re pouring billions into compute power, research teams, and top talent. For example, new hires are being offered access to 30,000 GPUs—the high-end chips that run modern AI systems—along with huge compensation packages.
From dropout to multimillionaire—in a single offer
Deitke, who dropped out of a PhD program not long ago, checks every box for a company like Meta. He’s worked on “multimodal AI,” the kind that can process images, text, and sound at once. That’s exactly the kind of model tech giants think will power the next leap in AI.
And they’re willing to put up NBA-style money to get people like him. Actually, scratch that. Better than NBA money.
Steph Curry recently signed a four-year deal worth $215 million with the Golden State Warriors. Deitke’s Meta offer? $250 million. That’s $35 million more than one of basketball’s biggest stars.
Why now, and why this much?
This moment feels different because it is.
AI has been through hype cycles before, but what’s happening now feels more like a gold rush. The talent pool is tiny, the stakes are sky-high, and the cash is flowing from companies with trillion-dollar valuations.
Unlike the Space Race or Cold War-era science, today’s AI race doesn’t have a clear finish line. There’s no “mission accomplished” moment. Whoever gets closest to building AGI could keep improving it endlessly. Imagine a machine that can outperform humans at almost everything—and even improve itself.
That’s the idea driving this surge. Whether it’s realistic or not is still up for debate.
According to one executive quoted by the Times: “If I’m Zuck and I’m spending $80 billion in one year on capital expenditures alone, is it worth kicking in another $5 billion or more to acquire a truly world-class team? The answer is obviously yes.”
Behind the scenes: chat groups, agents, and massive perks
Photo by Christina @ wocintechchat.com on Unsplash
Young researchers like Deitke aren’t just waiting around for offers—they’re actively negotiating. Informal agents, Slack and Discord groups, and even behind-the-scenes backchanneling help them maximize their deals.
It’s not just salary, either. Companies hand out computing resources, high-end GPUs, and access to key research infrastructure.
One co-founder of Vercept, Kiana Ehsani, even joked after Deitke’s offer went public: “We look forward to joining Matt on his private island next year.”
Final thoughts
The shift is clear: AI researchers aren’t just engineers anymore. They’re power players.
The last time we saw anything like this level of bidding power was the Gilded Age—where extreme industrial wealth met a very narrow band of expertise. Except now, the stakes may be higher, and the number of people who can actually steer this emerging technology is even smaller.
So yeah, a 24-year-old just got offered more money than the entire Apollo engineering team saw in years of work. Whether that’s brilliant or bonkers probably depends on what comes next in the AI race.
Keywords: Matt Deitke Meta salary, AI researcher compensation, AGI race, artificial superintelligence, AI talent competition, Meta AI hiring, multimodal AI expert, high paying tech jobs
