Nvidia CEO tells employees to us more AI, don't worry about their jobs

Even the company selling the world on automation can’t always get its own people to hit “go.” After Nvidia’s recent record earnings, an internal all-hands meeting — part pep-talk and part culture intervention — reportedly turned sharp when CEO Jensen Huang responded to an employee question about managers telling people to use AI less by demanding automation anywhere the technology can reach.

“My understanding is Nvidia has some managers who are telling their people to use less AI,” he reportedly said at the meeting, according to Business Insider, which was able to listen to it. “Are you insane?” He reportedly told employees, “I want every task that is possible to be automated with artificial intelligence to be automated with artificial intelligence” and told staff to use AI even if it doesn’t work yet — to use tools “until it does” and “jump in and help make it better.”

In the meeting, Huang addressed what Business Insider described as pockets of internal pushback.

Huang framed AI as the baseline skill that keeps people relevant as the company scales, not as a threat, reportedly reassuring the room, “I promise you, you will have work to do,” and pointing to the company’s aggressive hiring streak. Nvidia’s workforce has climbed from roughly 29,600 employees at the end of fiscal 2024 to about 36,000 a year later, and Huang reportedly said in the all-hands meeting that Nvidia is still “probably still about 10,000 short” of where it needs to be.

Nvidia isn’t alone in its AI push. Companies such as Microsoft, Meta, Google, and Amazon have all been nudging employees toward internal copilots and code assistants — and, in some cases, plan to factor AI use into companywide performance reviews. But unlike Nvidia, most other major tech firms have spent the past two years trimming their headcount or slowing their hiring. Nvidia, though, keeps swallowing data-center demand fast enough to justify adding thousands of people.

Some departments of Nvidia have reportedly made the jump to increased AI use. Engineers are working with an AI coding assistant called Cursor, and Huang in the meeting pointed to that as proof of how fast AI can reshape day-to-day work when it’s actually used — and used properly.

In a podcast interview last October, Huang said, “I’m hoping that Nvidia someday will be a 50,000-employee company with 100 million AI assistants in every single group.” Huang is reportedly a heavy chatbot user — using it as a “daily tutor” — and has said previously that he uses OpenAI’s ChatGPT daily in addition to tapping Gemini for more technical work, Grok for creative/artistic stuff, and Perplexity for fast research. Nvidia has formal partnerships with all but xAI’s Grok, which still uses Nvidia’s GPU at a serious clip. Huang’s company just signed a deal with Claude-maker Anthropic, so that chatbot might make its way into that rotation before too long.

Beyond his own campus, Huang has been sketching a much larger rewrite of work. On CNN, he said he is “certain [that] 100% of everybody’s jobs will be changed” because of AI, stressing that the tasks shift even if the business card does not. “The work will change, but it’s very likely … I’m still doing my job,” he said, arguing that history shows employment climbs alongside productivity when the economy keeps finding new things to build. “If the world runs out of ideas, then productivity gains translate to job loss.” AI, in his telling, is “the greatest technology equalizer we have ever seen” because “it empowers people. It lifts people.”

That seems to be part of a counterprogramming campaign against the current anxiety cycle — from IMF warnings to federal AI guidance — and Huang has been leaning into this type of comment more and more each quarter.

“You’re not going to lose your job to AI,” the CEO said earlier this year, “you’re going to lose your job to someone who uses AI,” treating AI fluency like table stakes in a labor market moving faster than the companies inside it. Automate the parts of the job that can be handed to a model, he appears to be saying, and there will still be plenty left that require human judgment, coordination, and whatever tasks appear once the old ones get uploaded into the mainframe. He then echoed those comments last month at Nvidia’s GTC in Washington, where he delivered the keynote address. The recent all-hands simply brought that message home to the people sitting closest to the GPUs.

So for now, the CEO who is selling the future of AI for a living is suddenly wrestling with an old-fashioned problem at home: human reluctance. Getting 36,000 people to retrain their muscle memory may turn out to be one of the hardest engineering problems on Huang’s list.