NVIDIA’s Jensen Huang Floats ‘AI Token’ Pay Plan as Compute Becomes Silicon Valley Recruiting Tool

NVIDIA CEO Jensen Huang believes generative AI has made computing power a direct driver of engineers’ output, meaning companies can no longer compete for talent solely through base pay, bonuses and equity. As model inference is priced by the token, access to AI compute could become a “fourth component” of compensation and influence Silicon Valley companies’ ability to attract top talent.

At NVIDIA GTC in March 2026, Huang proposed an “AI Token” compensation model that would give engineers an additional compute budget worth about half their salary, or 50% of annual pay, on top of existing compensation. The benefit would not be paid in cash. Instead, employees could use the allocation to access models and agent tools, improving software development and research efficiency.