For most of the past decade, the wealthiest person in any room at a frontier AI lab was almost certainly an investor. That assumption no longer holds. In the space of two years, a bidding war for a few hundred people has produced individual pay packages that would make a mid-cap hedge fund founder blush, and it has done so for employees who own no fund, carry no book, and answer to a manager.

The catalyst is Meta's aggressive push into what it now calls superintelligence. To assemble the team, Mark Zuckerberg has been writing offers that read less like compensation and more like acquisition prices for a single human being. Reporting collected by Euronews traces packages that run into hundreds of millions per person. The most extreme number sits higher still: the Wall Street Journal reported, via outlets including TechCrunch and CTech, that one researcher — Andrew Tulloch of Thinking Machines Lab — received an offer of up to $1.5 billion over at least six years, a figure Meta disputes as "inaccurate and ridiculous" and one that was initially rejected. Separate reporting elsewhere has pegged the packages for a ChatGPT co-creator at around $300 million to switch teams. Meta has publicly disputed some of the wilder numbers, and several remain unverified. The direction of travel, however, is not in dispute.
The economics of a few hundred people
The math is brutally simple. There are, by most credible estimates, only a few hundred people on earth who have actually built frontier systems at scale. That is not a labor market; it is a collectors' market. When Meta, OpenAI, Google DeepMind, Anthropic, xAI, and a clutch of well-funded startups such as Safe Superintelligence and Thinking Machines all want the same forty résumés, price discovery does what it always does with genuine scarcity. It goes vertical.
Structurally, these deals borrow more from professional sport than from Silicon Valley tradition. Alexandr Wang, brought in to help lead Meta's effort, is reported to sit on a package built around a modest base but anchored by roughly $100 million to $150 million in equity vesting over five years. That is the "franchise athlete" template: a signing structure, a vesting cliff, and a walk-away number large enough to make loyalty a rounding error.
This is not a labor market. It is a collectors' market, and price discovery does what it always does with genuine scarcity.
Richer than the people who used to finance them
Consider the comparison that should unsettle every allocator reading this. Demis Hassabis, who runs Google DeepMind, is estimated to be worth around $600 million, and reportedly collected a $3 million bonus tied to the Gemini program. Ilya Sutskever's stake in the ventures he has touched is measured in the billions; Greg Brockman's paper wealth at OpenAI has been pegged near thirty. These are not founders in the traditional equity-heavy sense. Several are, functionally, employees, engineers, and researchers who have been repriced overnight from well-paid professionals into the genuinely rich.
For the reader who has spent a career watching carried interest and fund economics generate two-comma net worths over fifteen-year horizons, the speed is the shocking part. A researcher can now out-earn a seasoned portfolio manager not across a career but across a single contract cycle, and do it without raising a dollar, without a lockup on a fund, and without the operational drag of running anything at all. The scarce asset is the person.
There is a caution buried in the euphoria. Franchise-athlete pay assumes the franchise keeps winning; the same equity that mints a fortune can evaporate if a lab's models fall behind or its valuation resets. But even a haircut on these numbers leaves a new caste, small, mobile, and extravagantly compensated, sitting above much of the finance industry that once assumed it owned the top of the tech food chain.
What separates the researcher who fields the $300 million call from the equally brilliant one who does not is rarely raw talent. It is proximity, who vouched for them, whose Signal thread they were on, which dinner they happened to attend. In a market this thin, the whole game is being in the room before the number is written down.
The room is the whole point.
The 1% is the private, verified network behind The 1% Journal — where members reach the people who actually move capital. Membership from $999.