Government officials and technology leaders called for stronger controls over artificial intelligence this week as incidents involving rogue agents continued, while President Trump said regulation was unnecessary. Several states are suing OpenAI over alleged harms to children. About 400 AI researchers raised cybersecurity concerns and called for global oversight, and Microsoft President Brad Smith said AI is less regulated than a carton of milk.

A central question is who should determine whether an AI system is safe. Dave Vellante argued that evidence and engineering for control should advance alongside model capabilities. He said policymakers need an independent standard with consequences and businesses should grant systems more authority only when deployments earn trust. Intellyx’s Jason Bloomberg separately argued that AI agents may not suit some companies.

Investors also showed concern that AI revenue growth may slow. Anthropic told investors its revenue rose 45% sequentially to $16.8 billion in the third quarter, a slower pace that could reduce its planned public offering valuation from $2 trillion to $1.5 trillion. Concerns about CoreWeave’s debt load weighed on neocloud stocks and their suppliers this week. Intel and Broadcom CEOs said capital is becoming a constraint on AI expansion as interest rates rise.

Despite those concerns, Nvidia, Meta and startups such as Reflection AI continued developing models. Companies are increasingly offering open-source or open-weight models as enterprises seek to customize models for their businesses. At the Open Compute Project Summit in San Jose this week, Cisco Systems, Nvidia, Intel and other suppliers introduced technologies intended to help AI scale up, out and across faster. SiliconANGLE said it plans to cover the open-model market soon. The article also noted that enterprise technology earnings season begins next week, with Netflix, Tesla, IBM, Meta and SAP reporting before Alphabet and Microsoft the following week. It listed events including Cloudflare Connect on October 19–21, the PyTorch Conference on October 20–21 and Oracle AI World on October 25–28. Earlier in the week, Anthropic’s third-quarter revenue growth, several states’ OpenAI lawsuits and calls for global AI oversight featured among its reported enterprise technology developments.

Government officials and technology leaders called for stronger controls over artificial intelligence this week as incidents involving rogue agents continued, while President Trump said regulation was unnecessary. Several states are suing OpenAI over alleged harms to children. About 400 AI researchers raised cybersecurity concerns and called for global oversight, and Microsoft President Brad Smith said AI is less regulated than a carton of milk.

A central question is who should determine whether an AI system is safe. Dave Vellante argued that evidence and engineering for control should advance alongside model capabilities. He said policymakers need an independent standard with consequences and businesses should grant systems more authority only when deployments earn trust. Intellyx’s Jason Bloomberg separately argued that AI agents may not suit some companies.

Investors also showed concern that AI revenue growth may slow. Anthropic told investors its revenue rose 45% sequentially to $16.8 billion in the third quarter, a slower pace that could reduce its planned public offering valuation from $2 trillion to $1.5 trillion. Concerns about CoreWeave’s debt load weighed on neocloud stocks and their suppliers this week. Intel and Broadcom CEOs said capital is becoming a constraint on AI expansion as interest rates rise.

Despite those concerns, Nvidia, Meta and startups such as Reflection AI continued developing models. Companies are increasingly offering open-source or open-weight models as enterprises seek to customize models for their businesses. At the Open Compute Project Summit in San Jose this week, Cisco Systems, Nvidia, Intel and other suppliers introduced technologies intended to help AI scale up, out and across faster. SiliconANGLE said it plans to cover the open-model market soon. The article also noted that enterprise technology earnings season begins next week, with Netflix, Tesla, IBM, Meta and SAP reporting before Alphabet and Microsoft the following week. It listed events including Cloudflare Connect on October 19–21, the PyTorch Conference on October 20–21 and Oracle AI World on October 25–28. Earlier in the week, Anthropic’s third-quarter revenue growth, several states’ OpenAI lawsuits and calls for global AI oversight featured among its reported enterprise technology developments.