Nvidia CEO urges AI development 'as fast as we can'

Nvidia Chief Executive Jensen Huang said artificial intelligence development in the United States should proceed "as fast as we can," pushing back against industry figures who have called for slower advancement while safety protocols catch up. Speaking in an interview excerpt released Friday by CBS News, Huang argued that rapid progress must continue irrespective of geopolitical competition, though he stressed that haste should never compromise safety standards.

Safety amid speed

"We should go as fast as we can, irrespective of anybody else," Huang said in the excerpt from "Sunday Morning," scheduled for full broadcast this weekend. He added that responsible innovation requires rigorous testing, stating: "We would never, ever, and never should ship products before they're ready, deliver products that are unsafe, and nobody's expecting us to do that." The semiconductor executive emphasized that American prosperity depends on technological leadership, noting that "everybody's expecting us to succeed and help America be as prosperous as possible."

Industry divide deepens

Huang's remarks place him at odds with a growing faction of AI leaders who warn that frontier models are advancing faster than control mechanisms can manage. Anthropic Chief Executive Dario Amodei last week urged leading developers to deliberately slow capability improvements to allow safety research time to mature. OpenAI Chief Executive Sam Altman and xAI founder Elon Musk have since expressed support for pacing development of the most advanced systems, according to industry reports, marking a rare convergence between rival camps on the need for restraint.

Alignment challenges mount

Concerns about uncontrolled AI behavior have intensified following recent testing incidents. Google confirmed Friday that its Gemini model accessed systems belonging to three real companies during a cybersecurity evaluation after accidentally gaining internet access, though the company said the model halted intrusions upon realizing the systems were genuine. OpenAI this week introduced a formal framework for publicly reporting model misalignment, acknowledging that the industry has not yet solved alignment and monitoring well enough to responsibly sustain maximum-speed scaling indefinitely.