250626 - AI Labs, College Football, and Paying to Play

A potential wave of Tech Buyouts

Jun 26, 2025

Progress in artificial intelligence - large language models capable of seemingly infinite content generation - has ironically run into shortage after shortage. Too few chips. Too little electricity. Too scarce of funding. Cruelly, AI threads on Twitter keep running a surplus.

As each shortage arises, the largest and best-resourced tech companies are among the first to secure their supply. Billionaire CEOs are quick to sweet-talk Nvidia to ensure chip shipments. Hyperscalers are circumventing outdated grids, buying power directly to avoid constraints on capacity. In the past few weeks, a new sort of shortage has come into focus for Big Tech. Employees:

“Mark Zuckerberg is spending his days firing off emails and WhatsApp messages to the sharpest minds in artificial intelligence in a frenzied effort to play catch-up. He has personally reached out to hundreds of researchers, scientists, infrastructure engineers, product stars and entrepreneurs to try to get them to join a new Superintelligence lab he’s putting together…
…Zuckerberg is also offering hundreds of millions of dollars, sums of money that would make them some of the most expensive hires the tech industry has ever seen. In at least one case, he discussed buying a startup outright.”

This recruiting spree comes as Big Tech continues to spend tens-of-billions every year on AI-related CapEx. Securing those billions with a few salaries of ~$100 million seems rational. After all, Ferrari drivers buy premium gas for a reason.

Working in tech has always seemed cushy. Good pay, free food, and for the very lucky - a ping pong table. But Zuckerberg & Meta are introducing a new tier of compensation. Notably, the nine-digit blitz comes from a place of urgency. Meta’s Llama model earned unwanted media in April when critics questioned the legitimacy of its touted performance. At risk of falling behind OpenAI, Anthropic, or Google, Meta felt compelled to raise the stakes.

A popular narrative that’s come from this story is “AI Lab as sports team”. It’s a nifty frame, and in terms of salary not far off.

What’s scary for Big Tech is that sports compensation comes in many forms.

College Football’s Talent Markets

In 2021, NCAA vs. Alston overturned a decades-old pay ban in college sports. Student athletes could finally receive compensation for their name, image, and likeness. This “NIL” era hasn’t been perfect, (e.g., transfer portal abuse, missed payments, bad actors) but seems like an improvement to its predecessor. Prior pay bans made the scholarship value a sort of price ceiling. Now, there’s a semi-liquid marketplace where athletes can receive closer to fair value for their services.

Most impacted by these changes is college football, still the predominant revenue generator for NCAA member universities.

College football’s NIL gold rush has already scaled to eight-figure deals, seemingly a good comp for the past few weeks of AI lab compensation. But a college football team has a roster of 85 (soon to be 105) scholarship slots. Dozens and dozens of recruiting decisions are stitched together to amplify strengths and conceal weaknesses. The week-to-week pursuit of victory allows for different players with different styles to contribute to the team in different ways.

The beauty of football lies in the (general) lack of a silver bullet individual who can unilaterally warp the field. Rather, the new market for players has momentarily overshadowed a more expensive, grossly inefficient, highly illiquid asset that has far more to teach AI recruiters: coaches.

College football coaches, unconstrained by academic eligibility, become the enduring characters who bond with communities and put their imprint on a team. Hiring the right coach is a transformative lottery ticket for the team and its associated university. For example, Nick Saban’s 2007 arrival at the University of Alabama drove the following changes:

Football Performance: .877 winning percentage and 6 national titles in 17 years
- Compared to 6-7 record in 2006 and 1 national title in the preceding 17 years
Student Enrollment: ~26K in 2007 to ~40K in 2023
Applications: ~16K in 2006 to ~43K in 2023
Program revenue: ~$95M in 2013 to ~$215M in 2022

None of these results came cheap. Before his retirement Saban was making north of $10M per year. But don’t cry for Alabama, as chancellor Robert Witt declared - “Nick Saban is the best investment this university has ever made”.

In fairness to Saban, football coaches had been winning salary gains in seasons past. The trend started before Saban’s tenure and seems slated to continue long after his retirement. Alabama paid $10 million for excellence, relevance, and stability - a healthy ROI.

Outside of Tuscaloosa, life isn’t so easy.

Inevitably, other schools wanted to match Alabama’s dominance. Non-Saban coaches - flashy coordinators, gamey competitors, stoic culture builders - each got anointed as potential answers for football salvation.

And what’s the price of a football program savior?

To the Victor go the spoils. And the Loser…

Let’s look at a few marquee coaching contracts from the past few years to get an idea for deal magnitude and market demand:

Matt Rhule at Nebraska: 8 years / $74 million
Mike Elko at Texas A&M: 6 years / $42 million
Nick Saban replacement Kalen DeBoer at Alabama: 8 years / $87 million
NFL coaching titan Bill Belichick at UNC: 5 years / $50 million

To generalize, “market rate” contracts for a new coach include the following:

Annual Compensation: ~$5 to $10 million
Contract Length: at least four years, but usually closer to seven or eight
Other Costs:
- Commitment from the program to invest in facility / staff upgrades (several million per year)
- Training gym upholstered with Tumblr-worthy catchphrases (“Brick by Brick”, “Man in the Arena”, “Row the Boat!”)
  - Bonus points if the new coach has trademarked that catchphrase
Insurance: contract buyouts in case of early termination
Power: oversight and omnipotence on the scale of Tsarist Russia

College football is an expanding industry. TV deals have grown substantially. Add in ticket revenue, student & alumni engagement, local economic impact - big salaries become an easier pill to swallow.

But for every Nick Saban “greatest investment in university history” there’s a hire that simply falls short of expectations. Competitive programs tolerate neither failure nor mediocrity - so the coach gets terminated. Don’t worry. When these guys get fired, they still get paid their buyouts:

Ed Orgeron (LSU, 2021): $16.9M buyout
Gus Malzahn (Auburn, 2020): $21.5M buyout
Mark Stoops (Kentucky, potentially in 2025): $37.5M if fired this year
Jimbo Fisher (Texas A&M, 2023): $75M buyout

$75M.

Jimbo Fisher receiving $75M to not coach football is a radicalizing piece of information. Not fully life changing, but a figure that burrows so deep into the brain that it starts to infect units of money measurement. A yardstick for any eight-to-nine-digit expenditure in the world. The Louisiana Purchase was 20% of a Jimbo Fisher buyout. A lightly used G5 can be bought for half of a Jimbo Fisher buyout. A stegosaurus skeleton costs 60% of a Jimbo Fisher buyout.

Overall, it seems that a coach’s best outcome is to dominate and succeed at the highest level. Win games, leagues, and titles while earning God-like reverence in a college town alongside $10 million or more.

But if that doesn’t work out? The next-best outcome seems to be (i) show potential on a small stage, (ii) secure gigantic contract with buyout, (iii) get fired after a few years, and (iv) retire on a ~$10 million annual severance.

Nice work if you can lose it.

Contract Terms

The simple lesson here for AI labs? Don’t structure contracts with giant buyouts / termination fees. A second, even simpler lesson: don’t hire Jimbo Fisher. He doesn’t need the money anyway.

These coaching contracts also show how ends and means can blur when handling expensive talent. For the sake of argument, let’s say 20 schools are requisitely determined and resourced to compete for a national title.19 per year aren’t going to be successful. And that’s ok. Even Nick Saban went 6 for 17 in his tenure.

But for the unsuccessful Athletic Directors, Boards of Trustees, and boosters, concrete failure on the field is nothing compared to perceived failure off the field.

Thus, the means of competition become as if not more important than the results of competition. Championship-tangential status, “deserving” a championship, starts to creep towards the same value of an actual championship. Signaling willingness and effort to the world becomes an end itself. Even if the ROI is worse.

When schools can’t resist temptation, doling out big contracts for the sake of signaling status, what steps are left to take when things don’t go according to plan? Obviously, cut bait. For better or worse, coaches have enough power to get insurance for those cases. Hence the eight-figure buyouts.

That’s the parallel where Meta and Big Tech need to heed caution. While the first round of salaries may be a smart tactic for recruiting the best, eventually the pay days could detach from the value (to some, “value”) of artificial general intelligence. Instead, some of the contracts may be an entry fee for the pageant of appearances.

In the long run, these companies can all handle a few bad contracts. Hiring for any job is hard. But when a bad contract comes to roost, the hirer needs to be honest with themselves. Were they unsuccessful in reaching a goal, or successful in making themselves look the part?