Jensen Huang: a $500,000 engineer should spend at least $250,000 a year on AI tokens

Nvidia CEO Jensen Huang described his own test for gauging an engineer's value to the company on the All-In podcast after GTC 2026, and the test comes with a concrete AI token budget.

What Huang's test involves

According to Huang, if an engineer with a $500,000 salary consumes less than $250,000 in AI tokens a year, which is less than half their pay, it "deeply troubles" him.

"If an engineer with a $500,000 salary consumes less than $250,000 in AI tokens a year, I will be deeply alarmed,"

Jensen Huang said on the All-In podcast.

• The statement was made on the All-In podcast right after GTC 2026 closed
• The example engineer's salary in Huang's test is $500,000 a year
• The minimum expected token spend is $250,000, or half the salary
• The test's author is Nvidia CEO Jensen Huang

Why this matters for the AI labor market

AI News titled its piece "Companies traded people for tokens, and the payoff is not yet visible," which points to the broader context of the remark: companies are rapidly increasing spending on AI tools for employees, expecting them to replace part of human labor or multiply its efficiency, but no measurable return on these investments is visible yet. Huang's position is, in effect, a public demand that Nvidia engineers use AI tools heavily, at a level comparable to a substantial share of their salary, or their value as employees is called into question.

What this means

A statement from the head of one of the world's most valuable technology companies formalizes a new engineer performance metric: not just code and results, but also how much an engineer spends on AI tools in dollar terms. If this approach spreads beyond Nvidia, it may change how companies assess employee productivity and plan AI infrastructure budgets: tokens turn from a line item into a measurable indicator of how deeply staff engage with AI.