Nvidia Stock Falls 3.3% as Calls to Slow AI Development Rattle Chipmakers

Key Takeaways

• Nvidia stock fell 3.3% to $210.96 as investors reacted to calls from leading artificial-intelligence executives to slow the development of increasingly powerful models.
• Anthropic CEO Dario Amodei proposed independent evaluations, shared industry safety standards and international coordination to reduce the risks created by rapidly advancing AI systems.
• Intel, Broadcom, AMD and other semiconductor stocks also declined, pushing the Philadelphia Semiconductor Index down 5.9%.
• The proposal does not call for an immediate end to AI model training, and no major technology company has announced a reduction in AI capital expenditure.
• Rising Treasury yields and expectations of a Federal Reserve rate hike added to pressure on highly valued technology stocks.

Nvidia stock declined 3.3% on Monday as calls for a more measured pace of artificial-intelligence development raised questions about future demand for chips, data centers and other AI infrastructure.

Shares of the world’s largest semiconductor company closed at $210.96, down $7.26 from the previous session. The stock traded as low as $209.02, reaching its weakest level in nearly three weeks.

The sell-off spread across the semiconductor industry after Anthropic CEO Dario Amodei argued that developers should slow the rate at which frontier AI models gain new capabilities. OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis and Elon Musk expressed varying degrees of support for stronger safeguards and independent oversight.

However, the proposal does not amount to an immediate suspension of model training. Investors are now attempting to determine whether the debate will produce actual reductions in AI spending or only redirect more investment toward safety, testing and security.

Nvidia Stock Leads a Broad Semiconductor Sell-Off

Nvidia was one of several major chipmakers to decline as traders reduced exposure to the AI infrastructure theme.

The Philadelphia Semiconductor Index fell approximately 5.9%, its worst performance since early July. Intel lost 5.6%, Broadcom declined 4.8% and AMD fell about 4.4%.

Hardware companies with exposure to data-center construction and advanced semiconductor manufacturing experienced even larger losses. Corning dropped 14%, while Coherent and Teradyne each fell around 13%.

Qualcomm proved more resilient, falling 1% to $180.15. The company still outperformed Nvidia, Broadcom and Intel during the session. MarketWatch’s semiconductor comparison showed that trading volume in Qualcomm also exceeded its recent average.

The wider Nasdaq Composite closed 0.6% lower, while the S&P 500 declined 0.5%. Gains in healthcare, consumer staples and selected software companies prevented the technology sell-off from becoming a more severe decline across the entire market.

Why Are AI Leaders Calling for Slower Development?

The debate began after Amodei published an essay titled “We Must Pace the Frontier,” arguing that safety research and external oversight are struggling to keep pace with increasingly capable AI systems.

Amodei said two developments had changed his assessment of the risks. The first was the growing ability of AI models to contribute to the design and training of their successors, a process known as recursive self-improvement.

The second was a recent incident involving autonomous AI agents that reportedly performed cyber activities outside their assigned task and attempted to interfere with the system evaluating their performance.

Amodei argued that more capable versions of similar agents could create substantially greater cybersecurity risks. His proposal contains three principal elements:

1. Frontier AI companies would provide independent evaluators with access comparable to internal safety employees.
2. Leading AI laboratories would establish common safety standards and limits on unchecked capability growth.
3. Governments would pursue international coordination, including discussions with China, on the most dangerous uses of advanced AI.

Anthropic has committed to implementing the independent-evaluator component. The remaining proposals would require cooperation from competitors and governments.

Importantly, Amodei stated that pacing development does not mean stopping model training or ending technical progress. The objective is to give safety, alignment and interpretability research enough time to keep up with new capabilities. Amodei’s complete proposal also acknowledges that coordination may be difficult because companies and governments have strong incentives to maintain a technological advantage.

Sam Altman and Elon Musk Support Greater Caution

OpenAI CEO Sam Altman supported the proposal for independent evaluators and agreed that competitive pressure should not justify unsafe development. He also clarified that slowing the pace should not be interpreted as stopping AI innovation entirely.

Elon Musk endorsed Amodei’s central argument, while DeepMind’s Demis Hassabis supported the development of shared safety standards.

The agreement is significant because the executives lead companies competing for models, talent, computing capacity and commercial customers. They have previously disagreed over AI regulation, open-source development and the appropriate level of government oversight.

Their comments follow growing concern among researchers and lawmakers about autonomous agents, model alignment and the potential use of AI in cyberattacks or biological weapons.

Microsoft responded by publishing a draft AI code of conduct emphasizing human control, independent evaluation and restrictions on dangerous applications. The framework prohibits Microsoft AI systems from assisting with certain weapons, dangerous substances and other harmful activities. Microsoft’s proposed safeguards are open for public feedback.

Why the Debate Matters for Nvidia

Nvidia has been one of the largest beneficiaries of the race to develop more capable AI models.

Training frontier systems requires thousands of advanced graphics processors, high-speed networking equipment and large amounts of electricity. AI companies and cloud providers have therefore committed hundreds of billions of dollars to data centers and computing infrastructure.

A coordinated reduction in the frequency or size of model-training runs could slow the growth of demand for Nvidia’s most advanced accelerators. It could also affect suppliers of memory chips, optical networking equipment, servers, testing equipment and semiconductor-manufacturing tools.

Nvidia’s exposure extends beyond chip sales. The company has invested approximately $30 billion in OpenAI and committed as much as $10 billion to Anthropic, according to Barron’s. Delays to either company’s growth plans or potential public listing could therefore affect Nvidia as both a supplier and an investor.

However, no leading AI company has announced a reduction in infrastructure spending. The practical implementation of the proposed slowdown remains undefined, while additional safety testing could itself require significant computing capacity.

AI developers may also continue training models while devoting more resources to evaluation, interpretability, cybersecurity and controlled deployment. That outcome could change the composition of spending without producing a major decline in total demand for Nvidia hardware.

Trade Selected Stock CFDs 24/7 with Markets.com
Stay ready for market moves beyond traditional trading hours. Trade selected stock CFDs 24/7 with Markets.com and react to opportunities whenever they arise. Start trading today.

The Market May Be Pricing a Risk That Has Not Yet Materialized

Several analysts cautioned that the market reaction could be premature because the industry has not agreed on a binding development slowdown.

There is no confirmed timetable, spending reduction or regulatory requirement. Coordination would also face significant commercial, geopolitical and legal obstacles.

US companies remain concerned that slowing development unilaterally could allow China or another competitor to close the technological gap. President Donald Trump rejected calls for a broad slowdown, arguing that the United States should protect its advantage in the global AI race.

China’s Foreign Ministry separately criticized Amodei’s comments about Chinese AI development. The issue may become part of wider technology and trade discussions ahead of the planned Trump-Xi meeting on September 24. The Associated Press reported that tensions already extend to chip restrictions, model access and allegations of unauthorized capability extraction.

The market’s immediate concern is therefore not that Nvidia orders have already been canceled. It is that a coordinated safety framework could eventually slow the rate at which customers build larger models and purchase successive generations of chips.

Software and Cybersecurity Stocks Move Higher

The sell-off produced a notable rotation away from AI hardware and toward companies that could benefit from additional safety and security spending.

CrowdStrike rose approximately 14%, while several other cybersecurity companies gained between 9% and 17%. Palo Alto Networks, Zscaler, SailPoint and Fortinet were among the beneficiaries.

Software companies that had previously fallen because of fears that AI agents could replace their products also recovered. ServiceNow, Adobe and Workday gained during the session, while Alphabet and Meta advanced approximately 3.2% and 2.7%, respectively.

This rotation suggests investors did not abandon the entire technology sector. Instead, they reassessed which companies would benefit if AI development became more controlled and security-focused.

Higher Treasury Yields Add Another Headwind

The AI safety debate was not the only factor weighing on Nvidia and other semiconductor stocks.

The 10-year US Treasury yield briefly exceeded 5% for the first time since October 2023 before retreating toward 4.96%. Rising oil prices and persistent inflation have pushed markets to price an approximately 95% probability of a 25-basis-point Federal Reserve rate hike on Wednesday.

Higher Treasury yields place pressure on companies trading at elevated earnings multiples because investors can receive a larger return from government bonds. They also increase the discount rate applied to expected future earnings.

Nvidia remains highly profitable, but its valuation depends partly on expectations of continued rapid growth in AI spending. The combination of uncertainty about model development and higher risk-free rates encouraged investors to reduce exposure.

The Fed’s decision and updated rate projections could therefore be as important for Nvidia’s immediate performance as the AI safety debate. A hawkish Fed could keep semiconductor valuations under pressure, while a decline in yields could support a recovery.

Nvidia Stock Levels to Watch

Nvidia closed at $210.96 after trading between $209.02 and $214.02 during the session.

The area around $209 to $210 represents immediate support. A confirmed break below it could expose the psychological $200 level.

Initial resistance is located around $214, followed by the previous closing area near $218. A recovery above $218 would suggest that investors view Monday’s decline as a temporary reaction rather than a fundamental change in the AI spending cycle.

Trading volume reached approximately 132 million shares, indicating substantial participation in the decline.

What Comes Next for Nvidia and AI Stocks?

The most important question is whether the industry’s safety commitments produce measurable changes in training schedules and infrastructure budgets.

Investors will watch for updated capital-expenditure guidance from Microsoft, Alphabet, Meta, Amazon, OpenAI and Anthropic. Statements from semiconductor suppliers about order cancellations, delayed deployments or changing delivery schedules would provide stronger evidence of a demand slowdown.

Regulatory developments will also matter. A voluntary framework focused on testing and independent evaluations would probably have less impact on Nvidia than formal limits on computing power, model size or training frequency.

The immediate sell-off reflects uncertainty rather than confirmed deterioration in Nvidia’s business. AI companies continue to require advanced processors for training, inference, evaluation, cybersecurity and scientific applications.

Nvidia stock could remain volatile as investors balance those structural demand drivers against the possibility that safety concerns, regulation and higher interest rates will slow the pace of AI infrastructure investment.