Nvidia’s Huang pitches AI tokens on top of salary as agents reshape how humans work
Nvidia CEO Jensen Huang delivers the keynote speech at the GTC AI Conference on March 18, 2025 in San Jose, California.
Josh Edelson | Afp | Getty Images
The advantages of working in Silicon Valley have long included high salaries. A new incentive may now be offered to some engineers: artificial intelligence tokens.
Nvidia CEO Jensen Huang announced a new compensation model on Monday. Giving engineers a token budget In addition to their base salary, they are effectively paid to deploy AI agents as a productivity multiplier.
The tokens, or units of data, used by AI systems can be spent to run tools and automate tasks, becoming “one of the recruiting tools in Silicon Valley,” Huang said.
“[Engineers] “They will earn a base salary of several hundred thousand dollars a year.” Huang said at the chipmaker’s annual GPU Technology Conference.
“I’ll probably give them half that [their base pay] as tokens… because every engineer with access to tokens will be more productive.”
The speech signaled Huang’s broader vision of the workplace, where engineers oversee a fleet of AI vehicles that can autonomously complete complex, multi-step tasks with minimal user input.
This is a vision that Huang has been developing publicly. He told CNBC last month that Nvidia employees will one day work alongside hundreds of thousands of AI agents.
“I have 42,000 biological workers and I will have hundreds of thousands of digital workers,” he said.
The comments come as concerns grow that AI agents — software systems that can independently execute complex, multi-step tasks — will eliminate white-collar jobs.
One memory Howard Marks, founder of Oaktree Capital Management, warned investors that there has been “an incredible advancement in the capabilities of artificial intelligence” that now allows it to “act autonomously”; this is a distinctive point that determines its ability to replace human labor.
“This difference is what separates a $50 billion market from a multi-trillion dollar market,” the veteran investor said.
Goldman Sachs estimates that AI could potentially automate tasks that account for 25% of all working hours in the US; That’s enough to fuel fears of what some harshly call a “business apocalypse.”
The bank sees a 15 percent productivity increase from AI, which could lead to the displacement of 6 percent to 7 percent of jobs during the adoption period.
“If AI proves to displace more labor than previous technologies, the risks tilt toward greater displacement,” said Joseph Briggs, senior global economist at Goldman.
Citing a study by economist David Autor, Briggs said that about 60 percent of today’s workers work in occupations that did not exist in 1940, and suggested that artificial intelligence will make some roles obsolete while creating others that do not yet exist.
AI agents drive software demand
Huang has an optimistic view on the impact of AI agents on the software industry, describing it this way: “Instead of reducing demand for software, AI agents will become its greediest customers.
His logic goes like this: More AI agents means more demand on the underlying software infrastructure they run on—the programs, tools, and computing resources that power them.
“The number of C compilers that we use, the number of Python programs that we have, the number of examples is growing very, very quickly because the number of our agents that are using these tools is increasing,” he said.
Bruno Guicardi, president and founder of information technology company CI&T, described the shift as nothing short of a paradigm shift. “A new layer of abstraction is being created through agents,” he said.
“Now software engineers can ‘tell’ computers what to do, not in a programming language, but in plain English. What used to take months now takes a few days. And we’re seeing it start to accelerate from there.”
‘Talent paradox’
Concern about workforce displacement caused by AI has been difficult to contain as companies struggle to find skilled workers.
Lewis Garrad, career practice leader at consultancy Mercer Asia, said the job market was currently experiencing a “talent paradox”, with 98% of senior executives expecting AI to lead to headcount reductions in the next two years, while 54% cited talent shortages as the biggest macro challenge.
Approximately 65% of executives expect 11% to 30% of their workforce to be redeployed or reskilled due to AI by 2026, Garrad estimates.
Garrad added that entry-level jobs face the greatest risk as AI eliminates “stepping stone” tasks historically used to train new workers, further widening the skills gap at a time when demand for AI-literate workers is rising.
Andreas Welsch, founder of consultancy Intelligence Briefing and author of The Human Agentic AI Edge, said roles involving data analysis, document processing, information comparison and preparation of initial reports were at risk of being “first in line” for displacement.
Goldman’s Briggs also predicts that the gross unemployment rate will increase by about half a percentage point as the labor market moves into a new era, acknowledging that even in the most optimistic scenario the transition will not be smooth.
However, new jobs will emerge, Briggs said, emphasizing that technological change is always the main driver of job growth in the long term through the creation of new jobs.
Tens of millions of people now work in industries that were science fiction a generation ago, such as computing, the gig economy, e-commerce, content creation, and video games.
However, integrating AI capabilities into existing enterprise workflows may ultimately be more difficult than the technology itself. Intelligence Briefing’s Welsch noted that roughly 80 percent to 85 percent of AI projects since 2018 have failed; That’s a sobering statistic for an industry brimming with enthusiasm.
“It is undesirable to have hundreds of thousands of agents creating more problems than they solve,” he said.
