Nvidia CEO Jensen Huang is one of the many tech CEOs encouraging the larger tech workforce to adopt AI into their workflow, especially those who deal with code. While Huang continues to build his fortune by selling AI chips to data centers, he now advocates for greater use of AI by engineers. Speaking on the All-In podcast on March 19, Huang declared that those software engineers and AI researchers earning $500,000 a year in salary should be burning through at least $250,000 annually in AI tokens.
Huang’s statements, which spread on social media soon after the podcast went online, have caught the attention of engineers and AI critics all around. In the shorter clip from the podcast footage, Huang drew a sharp analogy between engineers and chip designers – he stated that engineers who skimp on AI tokens today are similar to computer chip designers who refuse to use CAD tools. He called it unacceptable, especially in an age where AI is progressing at a fast pace.
“If that $500,000 engineer did not consume at least $250,000 worth of tokens, I am going to be deeply alarmed. This is no different than a chip designer who says ‘I’m just going to use paper and pencil. I don’t think I’m going to need any CAD tools,” he stated in the podcast.
Huang also painted his vision of the upcoming agentic AI era, in which highly paid engineers stop writing every line of code themselves and direct fleets of autonomous AI agents to handle complex tasks, which he believes could increase productivity by ten times.
Nvidia treats AI tokens like office laptops
Huang revealed the company is budgeting AI tokens for its 38,000 engineers the same way it budgets laptops, i.e., treating them as essential productivity tools rather than optional extras. “We are moving to agentic AI,” he stated, explaining how humans now become orchestrators to AI agents executing all the coding at superhuman speeds.
Supporters like Box CEO Aaron Levie are on board with the idea, predicting that token budgets will soon expand across entire teams as AI companies race to stay competitive. He sees this as the natural evolution in scaling AI adoption beyond individual power users.
Critics bring up cost concerns, ROI questions
Not everyone, however, is on board with the idea of having token budgets in annual company expenses. Critics argue the massive expenses on AI tokens could raise serious questions about return on investment, with some comparing it to Nvidia’s past push for more GPU usage — expensive and not always justified. Many others have expressed concerns on social media and communities as to whether every high-earning engineer requires the six-figure token budgets. Some also wondered if this idea risks becoming another layer of corporate hype.
However, for Huang, it’s all about leaning hard on modern AI tools to get rid of the menial tasks while giving humans the space and time to think about more intelligent stuff.