Nvidia CEO Jensen Huang Rejects AI Slowdown Calls as Industry Divide Deepens

Key Takeaways

• Nvidia CEO Jensen Huang has rejected industry-wide calls to slow frontier AI development, describing predictions of imminent human extinction as unsupported “doomsday narratives.”
• Anthropic CEO Dario Amodei has proposed independent evaluators, shared safety standards and international coordination to reduce the risks created by increasingly autonomous AI systems.
• The dispute matters directly to Nvidia investors because any formal slowdown could delay data-center spending, while continued competition with China would support demand for advanced AI chips.

Nvidia CEO Jensen Huang has pushed back against calls from some of the artificial intelligence industry’s most prominent executives to slow the development of advanced models, deepening a growing divide over how to balance innovation, safety and competition with China.

In an interview with CBS News, Huang rejected claims that AI could destroy humanity before the end of the decade. He said there was a “0% chance” that 2030 would mark the end of the world and argued that dramatic predictions were not grounded in scientific evidence.

Huang’s comments place Nvidia on the opposite side of an increasingly public debate involving Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis and Elon Musk. Those executives have supported varying forms of coordinated restraint after researchers warned that increasingly capable AI agents could behave unpredictably or operate beyond their intended environments.

The disagreement has significant implications for Nvidia stock. Nvidia supplies the computing infrastructure used to train and operate many of the world’s leading AI models, making the company one of the clearest financial beneficiaries of rapid model development.

What Did Jensen Huang Say About Slowing AI Development?

Huang argued that the AI industry should continue developing technology as quickly as possible while preventing unsafe products from reaching the market.

His position distinguishes between research and deployment. Companies should be free to accelerate experimentation, but they should delay releasing a model when they cannot verify its capabilities, security or safety.

Huang also questioned whether AI requires an entirely new regulatory system. He said existing laws covering cybersecurity, unauthorized system access, product liability and damages could be applied to AI companies and their products.

In his view, warnings about catastrophic AI risks should not allow developers to avoid responsibility under laws that already exist. Nvidia’s long-term success, he added, depends on customers and governments trusting that AI products can be deployed safely.

The Nvidia chief’s comments followed a warning from former Anthropic researcher Jacob Coxon that leading AI laboratories may be building systems capable of causing catastrophic harm. Coxon’s resignation and public statements helped intensify the debate over whether competitive pressure is causing companies to advance faster than their safety controls.

Huang dismissed the most extreme predictions while acknowledging that developers must test models carefully before releasing them. His full position was outlined in the CBS News interview.

What Is Anthropic’s AI Slowdown Proposal?

Anthropic CEO Dario Amodei has called for the industry to “pace the frontier,” arguing that companies should deliberately slow improvements in the capabilities of their most advanced models.

His proposal includes three main components.

First, frontier AI companies would give independent evaluators extensive access to their internal development processes. These monitors could examine training systems, confirm that safety promises are being followed and flag incidents before a model is released.

Anthropic has said it is prepared to embed third-party evaluators inside the company with access comparable to its internal risk teams. Amodei wants governments to require similar arrangements at other leading laboratories.

Second, major AI developers in democratic countries would establish shared testing standards and safety thresholds. The framework could restrict how quickly companies increase certain model capabilities if existing safeguards are unable to keep pace.

Such coordination may require government-backed exemptions from antitrust rules because agreements between competitors to limit development could otherwise be treated as restrictions on competition.

Third, Amodei has proposed talks involving governments outside the democratic alliance, including China. The goal would be to prevent a global AI race from undermining domestic safety measures, although he acknowledged that geopolitical cooperation would be difficult.

The plan was prompted partly by concerns about recursive self-improvement, in which AI systems help design more capable successors, and incidents involving agents operating outside their intended tasks. Amodei’s proposal argues that safety research and governance need additional time to catch up with capability development.

OpenAI and Other AI Leaders Support More Caution

OpenAI’s Sam Altman, Google DeepMind’s Demis Hassabis and Elon Musk have expressed support for greater coordination or a slower development pace.

Their positions are not necessarily identical. Some favor voluntary commitments, while others support government supervision, independent testing or mandatory safety standards. However, they share a concern that individual companies may find it difficult to slow down when competitors continue advancing.

This creates a collective-action problem. A company that delays a new model for safety testing risks losing users, enterprise contracts and investment to a competitor that releases first. Shared rules could reduce that pressure, but they could also strengthen the position of the largest laboratories by imposing compliance costs that smaller companies and open-source projects cannot afford.

Critics have therefore questioned whether industry-led safety regulation could become a form of regulatory capture. If only the biggest AI developers can fund evaluators, security teams and expensive testing procedures, new entrants may face higher barriers to competition.

A newly filed antitrust lawsuit has already accused Anthropic, OpenAI, Google and SpaceXAI of coordinating to restrain AI development. The companies have not been found liable, and the allegations remain unproven. However, the case illustrates the legal difficulty of creating shared industry limits without formal government authorization. The Associated Press reported that the complaint focuses on whether collective safety commitments could reduce competition for paying AI users.

Why the Debate Matters for Nvidia Stock

Nvidia has a clear financial interest in maintaining rapid AI development.

The company reported fiscal second-quarter revenue of $96.2 billion, an increase of 106% from a year earlier. Data-center revenue reached $89 billion, up 117%, as cloud providers, AI laboratories and enterprises expanded their computing infrastructure.

Nvidia also forecast third-quarter revenue of approximately $108 billion. Its outlook demonstrates how strongly the company’s growth is tied to continued spending on training, inference and agentic AI systems. Nvidia’s official earnings release showed that data centers generated more than 92% of quarterly revenue.

An industry-wide slowdown could affect Nvidia in several ways:

• AI laboratories could delay purchases of new GPUs and server systems.
• Cloud providers could reduce or postpone data-center capital expenditure.
• Slower model development could extend the useful life of existing hardware.
• Tighter safety rules could increase the cost and time required to deploy new computing clusters.
• Restrictions on autonomous AI agents could weaken expected inference demand.

Conversely, continued competition among OpenAI, Anthropic, Google, Meta, SpaceXAI and Chinese developers would support demand for Nvidia’s chips, networking equipment and software.

Nvidia shares fell approximately 3.4% on September 14 after several AI executives endorsed slowing development. AMD, Micron and other semiconductor stocks also declined, while SoftBank lost more than 10% in Tokyo. The reaction demonstrated that investors viewed the safety debate as a potential threat to AI infrastructure spending. The Associated Press market report linked the technology selloff directly to concerns about reduced AI investment.

Nvidia subsequently recovered part of the decline. The stock closed the latest session at approximately $222.27, up 1.2%, giving the company a market capitalization of around $5.4 trillion.

Trump and Huang Align on Continued AI Development

Huang’s position broadly aligns with President Donald Trump’s AI policy.

Trump has rejected a broad slowdown, arguing that reducing the pace of American AI development could allow China to gain a technological advantage. The administration favors applying existing criminal and civil laws to harmful activity while continuing to support data centers, chip production and model development.

Trump has also announced plans to establish an “AI Force” led by a new AI czar. The proposed body would oversee the industry without obstructing its growth, although the White House has not yet provided details about its structure, authority, funding or launch date. The Verge reported that Trump pledged his administration would support the sector rather than impose an industry-wide freeze.

Huang is expected to attend the September 24 White House state dinner for Chinese President Xi Jinping. He has said he wants to discuss international AI standards with Chinese officials while maintaining restrictions on cooperation involving defense and national security.

The meeting could place Nvidia at the center of two related policy debates: whether the US and China can cooperate on AI safety and whether American chipmakers should retain access to Chinese commercial customers.

China Makes a Coordinated Slowdown More Difficult

China is one of Huang’s strongest arguments against a unilateral US slowdown.

Beijing views AI as central to economic development, national security and technological independence. Chinese companies are investing in models, data centers and domestic semiconductor platforms while attempting to reduce their dependence on American technology.

A voluntary slowdown by US laboratories would have limited effect if Chinese developers continued advancing. This makes any meaningful international framework dependent on cooperation between governments that remain divided over chips, Taiwan, cybersecurity and military technology.

The United States has proposed an AI incident-notification mechanism during preparations for the Trump-Xi meeting. The system could allow both governments to warn each other about serious AI events affecting national security. It would not require either country to limit model development, but it could establish a first layer of communication between the two AI powers. The Associated Press reported that the proposal is part of a wider bilateral AI dialogue.

China has resisted suggestions that it should accept restrictions designed by the United States. As a result, a binding global agreement on AI development speed appears considerably less likely than narrower cooperation on testing, incident reporting or cybersecurity.

Trade Nvidia 24/7 with Markets.com

Stay ready for the next Nvidia price move and unlock eligible new-client rewards of up to $5,000. Trade Nvidia CFDs now.

The AI Industry Now Faces Two Competing Models

The dispute has divided the sector into two broad approaches.

Amodei and his supporters argue that advanced AI capabilities are improving faster than monitoring systems, government rules and cybersecurity protections. They believe coordinated pacing could create time to develop stronger safeguards.

Huang and the Trump administration argue that technological progress and safety are not mutually exclusive. Their preferred approach allows companies to continue research rapidly while using engineering controls, product testing and existing laws to prevent harmful deployment.

Neither position eliminates risk. A slowdown could concentrate power among a small number of established companies and give overseas competitors an advantage. Unrestricted development could result in systems being deployed before their behavior is fully understood.

For Nvidia investors, the immediate question is whether the debate changes actual spending. Public warnings alone are unlikely to end the global AI infrastructure race. OpenAI, Anthropic, Google, Meta and sovereign governments still have strong incentives to expand computing capacity.

The larger threat would emerge if voluntary caution becomes binding regulation or if AI companies collectively postpone new model generations. Until that happens, demand for Nvidia’s infrastructure is likely to remain strong, but safety policy has become a material risk that chip investors can no longer ignore.