Jensen Huang’s 'AI tokens' pitch
Nvidia CEO Jensen Huang said engineers should be evaluated on how many 'AI tokens' they use — a framing that signals compute access could become a productivity metric and even part of future compensation packages (indiatoday.in).
At Nvidia’s GPU Technology Conference keynote in mid‑March 2026, Jensen Huang proposed giving engineers an annual AI “token” budget roughly equal to half their base pay as an add‑on to salary. (cnbc.com) Huang used a concrete example on the All‑In podcast, saying “if that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed.” (indiatoday.in) Industry press explained tokens as the basic units AI models consume and cited Anthropic’s Claude 4.6 Sonnet pricing at about $0.000003 per input token and $0.000015 per output token — amounts that equate to a few dollars per million tokens. (indiatoday.in) Huang framed token budgets as part of a future workplace where Nvidia’s 42,000 “biological” employees will be augmented by “hundreds of thousands” of digital AI agents that they deploy and manage. (cnbc.com) In the same GTC presentations Huang tied token-driven agent adoption to hardware demand, projecting heavy orders for new Blackwell and Vera Rubin accelerators and suggesting multiyear demand that analysts estimated could approach $1 trillion through 2027. (cnbc.com) Public reaction included commentary from other CEOs noting broader economic implications, with OpenAI’s Sam Altman reportedly saying tokens could one day serve as a form of universal basic income. (aol.com)
