Jensen Huang arrived at a critical moment in Washington's reckoning with artificial intelligence, and he had a message for policymakers: the real danger isn't killer robots or mass unemployment. It's fear itself.
The Nvidia CEO told Axios that doomism around AI risks has become a policy trap, one that could cost America its competitive edge while doing nothing to prevent genuine problems. His blunt assessment cuts across an increasingly fractured tech establishment, where some of his peers have warned lawmakers about catastrophic risks from advanced systems.
Huang was direct about the stakes. "The fact that this is going to be the end of humanity, it's complete nonsense," he said. "The fact that this is going to destroy half of the American jobs is complete nonsense." The real threat, he argued, comes from scaring workers and companies away from deploying the technology altogether.
He pressed policymakers to resist reflexive restrictions built on hypotheticals. Before crafting rules around AI, Huang said, Washington should avoid basing policy on "science fiction" scenarios and should seek input beyond "one or two" tech CEOs. He also suggested some companies are using safety concerns strategically to secure regulatory advantage.
The tension between these two camps reveals a deeper split in Silicon Valley. Anthropic and OpenAI have spent months lobbying Washington to treat advanced systems as serious national security concerns. Huang did not name either company, though Nvidia supplies chips to both and depends on their business. His pushback suggests the semiconductor industry sees aggressive caution as bad for everyone's bottom line, including eventually the safety-focused firms themselves.
But the calculus in Washington is messier than Huang's either-or framing suggests. The Trump administration is simultaneously pushing rapid AI adoption and treating certain advanced systems as national security threats, particularly those from China. This week alone, Treasury Secretary Scott Bessent threatened sanctions against Chinese firms for what he described as "industrial-scale distillation attacks." Michael Kratsios, the White House's top technology official, accused Chinese startup Moonshot AI of stealing Anthropic's technology to build its Kimi model.
Commerce officials are also investigating whether Nvidia violated export restrictions on its Blackwell chips.
Huang's unusual influence with the Trump administration gives his warnings real weight in policy circles. He has cultivated that relationship carefully, praising the president and his top team, calling them "incredible people" who share his commitment to American success. That goodwill has translated into concrete wins for Nvidia, including approval to resume exporting certain AI chips to China.
The real test of Huang's worldview is about to arrive. As Chinese AI models improve and capabilities across the board accelerate, Washington faces a fundamental choice: treat openness and speed as competitive advantages, or treat them as risks that must be contained. Huang is betting the Trump administration sides with openness. But the investigation into Nvidia's exports suggests the government may be trying to have both.
Author James Rodriguez: "Huang's right that panic is bad policy, but his framing leaves out the part where Chinese firms are openly copying American technology to build their own military-grade systems."
Comments