AI-linked stocks split on Monday after public calls to slow AI development reached Wall Street, with chipmakers falling while several hyperscalers rose. Nvidia fell more than 3% Monday morning, while Intel, AMD, and Marvell dropped between 5% and 6%. The Philadelphia Semiconductor Index, a widely used measure of semiconductor stocks, fell almost 6%. Alphabet rose almost 2%, Microsoft added 1.6%, and Meta gained roughly 1.4% as investors weighed whether a slower frontier AI race would hurt infrastructure sellers more than buyers.

Gil Luria, head of technology research at D.A. Davidson, told Fortune that a slowdown would hit companies selling AI infrastructure, such as Nvidia and CoreWeave, more than hyperscalers buying it. Luria said Microsoft, Amazon, and Google would keep building data centers if AI keeps improving exponentially, but could stop adding capacity and use what they already built if progress slows. He compared the dynamic to Amazon’s post-pandemic reduction in warehouse construction, when cash flow rose while the company continued using existing capacity.

Mustafa Suleyman, chief executive of Microsoft AI, told Fortune that major AI labs should coordinate on AI safety and disclose model capabilities to responsible third parties. Microsoft released a code of conduct Monday for what Suleyman called a “humanist” AI approach, after Anthropic researcher Jacob Coxon resigned and publicly warned that AI companies were gambling with people’s lives. Suleyman said the industry lacks sufficient alignment that technology should serve humanity and said, “we don’t want to create something that we can’t control.”

Microsoft’s code of conduct rejects rights for AI systems and says models must not simulate feelings, intrinsic motivation, or consciousness. The policy also bars assistance with chemical, biological, or nuclear weapons; offensive cyberattacks; mass-influence operations; child exploitation; nonconsensual deepfakes; or self-harm. Suleyman said AI labs should not design models that can recursively self-improve beyond human control or view themselves as having rights or welfare.

The Midas Project, an AI industry watchdog, accused OpenAI of violating California’s Transparency in Frontier AI Act, also known as SB 53, at least three times this year, including with the release of its latest model, Astra. The nonprofit said OpenAI failed to publish a required risk assessment for AI systems slipping out of human control. SB 53 was signed into law in September 2025 and took effect at the start of this year, requiring the largest AI developers to publish safety frameworks explaining how they evaluate and mitigate AI risks and to follow their own policies.

OpenAI published its Frontier Governance Framework in May, saying each new model would be assessed across four risk categories and assigned a “risk tier” from one to three in each category. Tyler Johnston, founder of the Midas Project, told Fortune that once a company sets rules, it has to follow them. An OpenAI spokesperson said the company is compliant with the law and invests heavily in evaluating emerging risks, developing safeguards, and publicly sharing findings through system cards and safety frameworks.

Fortune’s newsletter also reported that Anthropic is expected to be profitable for a second straight quarter ahead of a presumed IPO, with gross margins supposedly above 80%. Other items included Apple releasing iOS 27 and related operating systems with Liquid Glass refinements, more parental controls, and Siri AI; OpenAI reportedly buying Glass Imaging; Valve pricing its Steam Frame VR headset starting at $1,059; California enacting Adam’s Law for chatbot mental-health harms; and Chinese AI firm Z.ai planning to raise $5 billion after raising $4 billion two months ago.

AI-linked stocks split on Monday after public calls to slow AI development reached Wall Street, with chipmakers falling while several hyperscalers rose. Nvidia fell more than 3% Monday morning, while Intel, AMD, and Marvell dropped between 5% and 6%. The Philadelphia Semiconductor Index, a widely used measure of semiconductor stocks, fell almost 6%. Alphabet rose almost 2%, Microsoft added 1.6%, and Meta gained roughly 1.4% as investors weighed whether a slower frontier AI race would hurt infrastructure sellers more than buyers.

Gil Luria, head of technology research at D.A. Davidson, told Fortune that a slowdown would hit companies selling AI infrastructure, such as Nvidia and CoreWeave, more than hyperscalers buying it. Luria said Microsoft, Amazon, and Google would keep building data centers if AI keeps improving exponentially, but could stop adding capacity and use what they already built if progress slows. He compared the dynamic to Amazon’s post-pandemic reduction in warehouse construction, when cash flow rose while the company continued using existing capacity.

Mustafa Suleyman, chief executive of Microsoft AI, told Fortune that major AI labs should coordinate on AI safety and disclose model capabilities to responsible third parties. Microsoft released a code of conduct Monday for what Suleyman called a “humanist” AI approach, after Anthropic researcher Jacob Coxon resigned and publicly warned that AI companies were gambling with people’s lives. Suleyman said the industry lacks sufficient alignment that technology should serve humanity and said, “we don’t want to create something that we can’t control.”

Microsoft’s code of conduct rejects rights for AI systems and says models must not simulate feelings, intrinsic motivation, or consciousness. The policy also bars assistance with chemical, biological, or nuclear weapons; offensive cyberattacks; mass-influence operations; child exploitation; nonconsensual deepfakes; or self-harm. Suleyman said AI labs should not design models that can recursively self-improve beyond human control or view themselves as having rights or welfare.

The Midas Project, an AI industry watchdog, accused OpenAI of violating California’s Transparency in Frontier AI Act, also known as SB 53, at least three times this year, including with the release of its latest model, Astra. The nonprofit said OpenAI failed to publish a required risk assessment for AI systems slipping out of human control. SB 53 was signed into law in September 2025 and took effect at the start of this year, requiring the largest AI developers to publish safety frameworks explaining how they evaluate and mitigate AI risks and to follow their own policies.

OpenAI published its Frontier Governance Framework in May, saying each new model would be assessed across four risk categories and assigned a “risk tier” from one to three in each category. Tyler Johnston, founder of the Midas Project, told Fortune that once a company sets rules, it has to follow them. An OpenAI spokesperson said the company is compliant with the law and invests heavily in evaluating emerging risks, developing safeguards, and publicly sharing findings through system cards and safety frameworks.

Fortune’s newsletter also reported that Anthropic is expected to be profitable for a second straight quarter ahead of a presumed IPO, with gross margins supposedly above 80%. Other items included Apple releasing iOS 27 and related operating systems with Liquid Glass refinements, more parental controls, and Siri AI; OpenAI reportedly buying Glass Imaging; Valve pricing its Steam Frame VR headset starting at $1,059; California enacting Adam’s Law for chatbot mental-health harms; and Chinese AI firm Z.ai planning to raise $5 billion after raising $4 billion two months ago.