AI safety and financing worries multiply but everyone’s still racing full speed ahead
SiliconANGLE Robert Hof ● Covered by 5 sources
AI safety fights are piling up, and so are funding jitters. But companies are still shipping models and buying chips as fast as ever.
Based on reporting by SiliconANGLE, Robert Hof — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
The AI safety debate just got louder. Government officials want more regulation after weeks of incidents involving rogue agents, even as Trump says it isn’t needed. Several states are suing OpenAI over alleged harm to children. And inside tech, the complaints are getting harder to ignore: hundreds of AI researchers have raised cybersecurity concerns, and Microsoft president Brad Smith says AI is less regulated than a carton of milk.
The deeper problem is simple and ugly: who gets to declare an AI safe? Dave Vellante’s latest Breaking Analysis argues that the evidence and the engineering for control need to move in step with model capability. His prescription is blunt. Policymakers need an independent standard with consequences, and businesses should only hand over more authority when deployment has earned trust. Jason Bloomberg at Intellyx goes further, saying some companies may decide AI agents are not the answer at all.
Money is adding its own pressure. Investors are getting nervous that the AI revenue boom may not keep running at the same speed, especially for names like OpenAI and Anthropic. Anthropic disclosed 45% revenue growth from the second to the third quarter, but that was still a slowdown, and it could pull its planned public offering valuation down from $2 trillion to $1.5 trillion. CoreWeave’s debt load also rattled the market, pushing down shares of neoclouds and suppliers this week.
And yet the spending spree goes on. Intel and Broadcom’s CEOs warned that capital is already becoming a brake on AI expansion as interest rates creep up, but Nvidia, Meta and startups like Reflection AI are still pushing new models, including more open-source and open-weight releases aimed at enterprise customization. At the Open Compute Project’s summit in San Jose, Cisco, Nvidia, Intel and other suppliers showed off hardware meant to help AI scale faster. The race is still on, even if the bill is getting harder to ignore.
My take — AI-written commentary, not fact-checked reporting
The market is finally discovering that “move fast” is not a financing plan. AI companies want the upside of full speed and the paperwork of none, which is a charming strategy right up until the debt or the regulators arrive. Open models may be the enterprise favorite, but they also make the safety hand-waving look even sillier.
Read more about this at: SiliconANGLE