Jensen Huang has a message for Washington: stop letting Hollywood scripts write AI policy.
The Nvidia CEO sat down on July 23 for what amounted to a public lobbying session aimed squarely at US policymakers, arguing that exaggerated fears about artificial intelligence are doing more damage to American competitiveness than the technology itself ever could. His core thesis is simple: the people warning that AI will end civilization are drawing from science fiction, not science.
The case against panic-driven regulation
Huang didn’t mince words about the doomsday crowd.
“The fact that this is going to be the end of humanity, it’s complete nonsense.”
But Huang’s frustration wasn’t just philosophical. He specifically warned that the Trump administration might “overcorrect” on AI policy, creating a regulatory environment where competitors could weaponize safety rhetoric to gain advantages.
His timing wasn’t accidental. The same day Huang delivered his remarks, Treasury Secretary Scott Bessent was issuing threats of sanctions against Chinese AI models.
Huang argued that Washington should consult a wider range of stakeholders rather than relying on the perspectives of a handful of CEOs.
The trillion-dollar backdrop
Back in March 2026, he projected demand of roughly $1 trillion in chips by 2027. That reflects the insatiable appetite for compute power driven by AI training, inference, and deployment across virtually every industry.
Nvidia has been under heightened scrutiny over its chip exports to Chinese companies, with potential investigations hanging overhead. Meanwhile, Chinese firms like Moonshot AI have been releasing competitive open-source AI models, eroding the assumption that export controls alone can maintain US technological superiority.
Huang has consistently pushed for federally unified AI regulations rather than a patchwork of state-level rules. A coherent federal framework would give companies a single set of rules to follow.