When AI Costs More Than the Engineer

The core tension is simple: frontier AI labs burn compute at a rate that far exceeds their own payroll. Anthropic spends 2.

3x its payroll on compute — roughly $2m of compute per employee per year against a likely all-in comp of $500k+.

The rest of the software market trails dramatically: the top 1% of companies spend $89k per engineer on AI (40% of a fully-loaded $224k senior engineer salary), while the median spends just $137.

That two-order-of-magnitude gap between frontier and median defines the strategic challenge for builders who need to decide how aggressively to invest in AI infrastructure.