Definition
The global competition for the scarce engineers and researchers who build AI, where short supply drives pay sharply higher.
At a glance
- Demand dwarfs supply: roughly 1.6M open AI roles globally, and a 2026 survey of 39,000+ employers ranked AI skills the world’s hardest to hire.[1]
- Average AI engineer base pay hit about $206,000 in 2025, with a 56% wage premium just for AI skills.[2]
- Top researchers got athlete-sized offers: Meta up to $300M over four years.[3]
- Giants buy whole startups just for the staff, and regulators are now watching.[4]
Why pay is so high
Scarcity. Few people can build cutting-edge AI, yet nearly every large company wants them, so prices rise.[1] Engineer pay jumped ~$50,000 in a year, with generative-AI specialists earning 40-60% more on top, as U.S. demand keeps climbing.[6]
Buying the team, not the product
When individuals are too hard to recruit, big firms buy whole startups for their staff (an “acqui-hire”): Microsoft-Inflection ($650M), Google-Character.AI, Meta’s $14B Scale AI stake.[4] The FTC and DOJ now probe these as “pseudo-acquisitions” that may starve rivals of talent.[5]
Bottom line
For most owners, the move isn’t to win this auction but to route around it with vendors, tools, and contractors rather than competing for sky-priced AI staff.