Nvidia CEO rejects AI slowdown calls

At the All‑In Summit in Los Angeles, Nvidia (NVDA) CEO Jensen Huang argued that the United States can continue to innovate rapidly in artificial intelligence while still developing the technology safely. He said he does not agree that AI development itself must be slowed for safety reasons, directly pushing back on calls to reduce the pace of frontier AI research.

Huang criticized public warnings that AI could "wipe out humanity" or pose extinction‑level threats, stating that such claims lack scientific grounding. He described this form of doomerism as generating needless anxiety, and maintained that the focus should instead be on concrete safety practices built into how systems are engineered and deployed.

Rather than advocating for new laws or a development pause, Huang promoted a framework of industry and engineering controls to manage AI risks. He highlighted the role of sandboxes or pre‑release testing environments where models can be examined before broader deployment.

Huang also called for continuous monitoring of AI systems once they are in use, along with independent third‑party assessment bodies. He compared these organizations to financial auditors, suggesting they could test and audit models before public release to check for unsafe behaviors or flaws.

He argued that many recent problems, such as model jailbreaking and other risky behaviors, have largely occurred inside frontier research labs. In his view, shifting work from experimental labs into structured engineering and product development environments enables better processes, human oversight and testing to prevent similar issues from recurring.

During Huang’s onstage session at the summit, he received a surprise live phone call from U.S. President Donald Trump. On the call, Trump dismissed fears of an AI takeover as a "hoax," aligning with Huang’s rejection of apocalyptic narratives around the technology.

Trump also described data centres as "the oil of the next 20 to 25 years," emphasizing their growing role as critical economic infrastructure. Huang publicly received these remarks, which reinforced the event’s focus on AI as a driver of economic change and ongoing debate over how heavily it should be regulated.

Huang used the summit appearance to defend the practical use of open‑source AI models, describing them as important for AI sovereignty. He suggested that accessible, shared model foundations can help countries and organizations maintain control over their own AI capabilities.

He cautioned against broad measures that would restrict Chinese researchers and firms from accessing U.S. AI technology. Huang said that such sweeping limits risk ceding advantage, and instead argued for maintaining broad access while applying the kinds of engineering safeguards and oversight mechanisms he outlined for the industry.