Nvidia CEO Jensen Huang said Wednesday (June 11) that Anthropic CEO’s prediction of artificial intelligence (AI) destroying half of entry level white collar jobs “doesn’t make sense.”
Dario Amodei reportedly made the comment last month, adding that he expects it to happen within one to five years. He also believes that it would raise the unemployment rate to 10% to 20%. The unemployment rate for May was 4.2%, according to the Bureau of Labor Statistics.
During a private briefing with journalists and analysts at the VivaTech 2025 conference in Paris, a relaxed and candid Huang said it didn’t make sense to him that Amodei believes AI will take all these jobs and also imply that Anthropic should be the only one to develop AI.
Also, Huang pointed out why it is illogical for Amodei to say AI can be dangerous.
Huang said if AI was so dangerous, the solution is not to leave it in one company’s hands — like Anthropic — but rather open it to the world.
“Do it in the open,” Huang said. That way, there will be many eyes on it.
Expect AI Changes
Huang’s personal belief is AI will change everyone’s jobs, including his. Since AI makes companies more productive, he believes it will actually end up creating more jobs than cutting them.
Those companies that are cutting jobs are ones that have run out of good ideas, Huang said.
Asked if he believes AI poses an existential threat to humanity, Huang said it won’t happen if one rogue AI system is watched by many other AI systems. Right now, there isn’t a lot of monitoring in place. But as AI costs come down, there can be as many AI systems watching as possible.
The Future of AGI
As for when he thinks society can achieve artificial general intelligence, or AGI, Huang said there are engineering and theological definitions of AGI.
On an engineering level, the AI system has to pass math, coding, science and other tests. If the AI can match or surpass a human’s general abilities, then it has achieved AGI. However, Huang said he can’t answer the theological question.
On autonomous driving, Huang said large language models (LLMs) and generative AI will enable the capability. Today’s self-driving vehicles have bolted-on capabilities — radar, cameras, GPS, route planning, mapping solutions, connectivity and other modules. That makes them brittle, Huang said.
With LLMs, Nvidia is able to make an end-to-end solution, he added.
Huang also believes that one day, humans will be able to talk to their cars just like one converses with ChatGPT. Even the lawn mower could one day talk to their users.
The CEO also shed light on news regarding Mistral AI developing AI data centers in partnership with Nvidia.
The plan of Mistral AI CEO Arthur Mensch is to create the “CoreWeave” of Europe, Huang revealed.
CoreWeave provides AI data centers in the cloud. Today’s cloud computing giants offer cloud services using traditional data centers that are not optimized for AI.
Projects like OpenAI’s $500 billion Stargate aim to build AI data centers with intense computing power and energy sources needed to process AI workloads.
Mensch plans to create AI data centers that will offer cloud computing services to clients, just like CoreWeave does.
“That’s his plan,” Huang said. Nvidia’s aim is “to find and introduce him to financing and customers we work with here” as well as provide chips.
The session ended with Huang taking selfies with analysts and journalists in an intimate setting. “I’m super afraid of public speaking, you know.”
Read more: Oracle to Buy $40 Billion Worth of Nvidia Chips for First Stargate Data Center
Read more: Why Is Silicon Valley Spending a Fortune on AI Data Centers?
Read more: OpenAI Plans Grand Expansion While Losing Billions. Sound Familiar?
