Jensen Huang, CEO of the world's most valuable company by market value, described the computers of the future as “manufacturing equipment” that will generate “AI tokens” that can also be used for the employee market, reports Business Insider.
In Huang's context, tokens are units of text—a word or part of a word—that determine how AI activity is measured and charged. A short word can represent a single token, while longer words can be split into several. As a general rule, a token is about four characters long.
Large language models like OpenAI's ChatGPT or Anthropic's Claude track the number of tokens processed when a user enters text and how many are generated in response — and the AI giants bill their client companies accordingly.
The more someone works with a text, the more tokens are needed for the AI models to process it. Unlike existing software costs, which are billed as subscriptions or flat fees, AI is billed based on usage.
What AI tokens have to do with employee pay
Speaking at Nvidia's GTC AI Conference this week, Huang said that in the future, as AI technology becomes even more central to work, engineers could be given “token budgets” to help them become more productive.
In his keynote speech on Monday, he even suggested the idea of giving Nvidia engineers tokens worth half their annual salary to attract talent.
Huang echoed that idea on Tuesday. He said those costs will be worth it, especially for highly paid engineers who can achieve significant productivity gains with the help of so-called “AI agents,” that is, autonomous applications capable of performing various head-to-tail tasks.
“If I were to add another $100 a day to them as the cost of running the model — the cost of tokens — I would be more than willing to do it,” Huang said of paying engineers extra, especially during busy periods. “If I were to add even $1,000 during the peak periods, I'd be more than happy to do it,” he pointed out.
Huang added that more powerful and energy-efficient hardware can generate tokens at lower and lower costs over time.
Business Insider notes that Huang is not alone in promoting this idea. The concept is starting to take shape across the tech industry, with engineers asking about computing resources during interviews and executives considering them as part of compensation packages.
