Will extreme AI researcher compensation keep escalating, or normalize as the skill spreads?

Nick Frosst · published Sep 1, 2025

It makes sense that top AI researchers are rewarded extremely well

There genuinely are people bringing that much value into a hugely impactful industry, and the work is demanding, requiring experience, ingenuity and dedication

21:47 20VC: Cohere Founder on How Cohere Compete with OpenAI and Anthropic $BNs | Why Counties Should Fund Their Own Models & the Need for Model Sovereignty | How Sam Altman Has Done a Disservice to AI with Nick Frosst

Andrew Feldman · published Oct 6, 2025

Some engineers and scientists have capabilities that no quantity of other talented people working together can replicate, so paying extraordinary people extraordinary sums is never the mistake — paying mediocre people too much is

No company ever went bankrupt by overpaying truly extraordinary people; companies fail by overpaying mediocre ones — and society already pays huge sums to entertainers like Charlie Sheen whose net productivity is arguably lower

Scope: explicitly agnostic on whether they should out-earn soccer or basketball players

34:25 20VC: Cerebras CEO on Why Raise $1BN and Delay the IPO | NVIDIA Showing Signs They Are Worried About Growth | Concentration of Value in Mag7: Will the AI Train Come to a Halt | Can the US Supply the Energy for AI with Andrew Feldman

Harry Stebbings · published Feb 23, 2026

The $100M-$500M AI pay packages are justified

Very few people know how to do this and it is a very difficult job

61:02 20VC: Inside Coatue's $70BN Machine: Why Price Matters Least | Why Mega Markets are the Most Important | How to Assess Durability of Revenue and Margins in AI with Lucas Swisher

Joelle Pineau · published Nov 3, 2025 · hedged

Multibillion-dollar price tags for individual AI talent are not necessarily needed

31:21 Multibillion dollar individual packages are not necessary

Harry Stebbings · published Oct 6, 2025

Paying an AI chief scientist a billion dollars is economically rational if they add $50 billion of enterprise value

Pure economic rationale: compensation should track value generated

34:44 Pay should scale with marginal enterprise value generated

Andrew Ng · published Nov 17, 2025 · hedged

Whether the enormous AI pay packages are justified is very hard to say, but it is a good thing that money is flowing into paying AI people really well

18:39 Justification unknowable but money flowing to ai talent is healthy

Brendan Foody · published Jun 1, 2026 · hedged

Extreme AI researcher compensation will keep escalating for a shrinking elite group, but the broader market should normalize as more people acquire frontier training know-how and supply increases

As knowledge of how labs operate and how to train frontier models spreads, supply of that skill set grows and pricing becomes more reasonable

62:20 Top tier keeps escalating while the broader market normalizes

Jonathan Ross · published Sep 29, 2025

The war for talent is more aggressive than at any point in history, but only in tech

Sports salaries have long been insane; the difference is sports has a limited number of teams and salary caps while tech has unlimited startups competing for the same people

56:16 Tech talent war is uniquely fierce due to unlimited competing employers unlike capped industries

David Cahn · published Oct 27, 2025

The logic justifying huge pay packages — that a 1% increase in the probability of a trillion-dollar outcome is worth $10B — is abused because we systematically overestimate the individual's percentage contribution

Human brains reason badly at that scale of number; the true contribution could be a hundredth or ten thousandth of a percent, and broader macro variables may drive AI progress more than any individual

9:25 Individual contribution to outcomes is systematically overestimated so mega packages are unjustified