Pave and Nua Group released "The State of AI Talent: A Compensation & Workforce Report" on Aug. 24, drawing on compensation data from more than 9,000 companies, including 80% of the Forbes AI 50. The report shows that AI Research Scientists now command median new-hire equity grants above $4 million, while base salaries across AI job families remain relatively compressed. For HR teams, the findings signal that the real competition for AI talent happens in equity packages and at the point of hire, not in annual salary adjustments.
The report analyzes employees mapped to AI and ML jobs through automated connections to HRIS, ATS, and equity management systems. It identifies three distinct job families - AI Engineering, ML Engineering, and AI Research Scientist - each with its own pay curve, seniority mix, and hiring trajectory. That separation matters for how companies structure roles and set compensation bands.
Equity, not base salary, drives AI pay
Median base pay is relatively compressed across the three core job families. ML Engineering leads at senior levels, with $321K at P6 and $347K at M6. But median new-hire equity for AI Research Scientists reaches $4.09M at P6 and $4.72M at M6 - more than double their counterparts in ML Engineering and AI Engineering at the same levels.
At private San Francisco Bay Area companies that have raised $1B-$5B+, the 90th percentile new-hire equity packages for AI Research Scientists at the most senior levels reach into the $45M-$60M range. HR professionals benchmarking AI roles against base salary alone will miss where the market is actually moving.
AI Engineering is growing fastest, but talent stays concentrated
AI Engineering is now the fastest-growing job family. Its share of all employees grew from 0.010% in Q4 2023 to 0.185% in Q2 2026, and its share of new hires spiked from 0.11% to 0.35% over the same period - roughly 10x growth. Meanwhile, AI/ML talent remains concentrated at scale and on the coasts: just 12% of companies with 1-99 employees have any AI/ML talent, versus 91% of companies with 3,000+ employees. California alone holds 49.3% of US-based AI/ML professionals, and the US accounts for 57.3% of this talent globally.
Retention is becoming harder. Over the last 12 months, AI/ML turnover for individual contributors was 21.8%, versus 16.8% for software engineering at the same companies. Recent hires in AI/ML are also out-earning their incumbent peers, with median base salary premiums of up to 10% for AI Research Scientists hired in the past six months at the same level.
"AI talent is the fastest-moving hiring market we have ever measured, and it's moving in ways an annual survey simply cannot see," said Matt Schulman, CEO of Pave. "Base salaries look compressed on paper, but the real competition is happening in new-hire equity and at the point of hire, where recent hires are commanding meaningful premiums over incumbents. If you're benchmarking this market against data that's even six months old, you're not looking at the market - you're looking at its history."
A framework for matching pay structure to company type
The report introduces a framework developed by Nua Group that sorts organizations into three categories: AI Innovators, AI Integrators, and AI Implementers. Each has a distinct hiring profile and cash-and-equity posture. The goal is to help companies "know their AI lane" before building a pay structure.
"We built this framework because we kept seeing the same pattern across clients," said Ryland Bauer, Total Rewards Advisor at Nua Group. "A company would come to us needing help pricing 'AI Engineering talent,' and what they actually needed looked completely different depending on whether they were building frontier models, embedding AI into a product, or applying it inside an existing business. Once you know which of those three you are, the rest of the compensation decisions get a lot more straightforward."
"The most expensive mistake we see isn't paying too much - it's paying for the wrong role," Schulman continued. "The market has matured to the point where AI Engineering, ML Engineering, and AI Research Scientist are genuinely distinct job families with distinct pay curves. Companies that manage that job architecture right, and then revisit their bands more often than they would for any other engineering family, are the ones winning the best talent."
The full report, including level-by-level benchmarks and turnover analysis, is available at Pave's website. For HR professionals managing AI roles, the report's practical takeaways - treating AI job families as distinct, benchmarking equity rather than just salary, and revisiting bands more frequently - align with the AI for Human Resources resources available through Complete AI Training.
Why this matters for HR professionals
For HR teams, the practical takeaway is to separate AI job families in your compensation architecture and benchmark equity grants, not just base salary. The report shows that a company hiring an AI Research Scientist at a late-stage private company should expect to compete on equity packages that can reach eight figures, while an AI Implementer hiring for applied roles will face a different market entirely. The AI Learning Path for HR Managers covers the recruitment and compensation analysis skills needed to manage these decisions effectively.
Given the 21.8% turnover rate for AI/ML individual contributors, HR teams should also treat retention as an ongoing process rather than an annual review cycle. Recent hires commanding 10% base salary premiums over incumbents means pay compression is a real risk - and one that requires more frequent market checks than traditional engineering roles.
Your membership also unlocks:
