Former SEC Chair Gary Gensler told CNBC the AI trade is "unlikely to last," warning today's infrastructure buildout will eventually plateau and reverse, "it's just a question of when and how steep." His mechanism is specific: if AI companies stop beating quarterly earnings estimates and start flatlining, "the capital markets will reprice it," starting with memory chip suppliers losing pricing power first. Worth noting the timing lines up with real market movement, the PHLX Semiconductor Index dropped over 5% in the five days following Amodei's and Altman's calls for an AI slowdown, and AI capex is projected to exceed $795B in 2026, heading toward $1T in 2027. Gensler flagged private credit, vendor financing, and loan guarantees tied to that buildout as the areas to watch if the cycle actually turns, that's where the real fragility would show up first, not in any single company's stock price.

President Trump called Nvidia CEO Jensen Huang mid-session at the All-In Summit and got put on speakerphone in front of the crowd. His message: AI safety fears are a "hoax," data centers create real wealth, and AI is "bigger than the internet." He's argued the same point elsewhere, that the upside outweighs the risk, and that staying ahead of China matters more than slowing down. Huang's response was direct: "You're right, we're not going to let that happen, sir." Let's note the timing: this call landed the same week Dario Amodei published his essay calling for the industry to "pace the frontier," and Musk and Altman both publicly agreed with it. Trump's live rebuttal, delivered through Huang's own microphone, was a pointed answer to that exact moment, not a random aside.

Something genuinely unusual happened this weekend: Dario Amodei, Sam Altman, and Elon Musk, three executives who've spent years publicly clashing, converged on the same position. Amodei published an essay calling for AI firms to deliberately "pace the frontier," warning that capabilities are improving faster than researchers can understand or control. Musk responded within hours: "Dario is right." Altman followed, calling independent safety evaluators "a great idea" and confirming OpenAI would grant them "employee-like access." Amodei's framework has three stages: third-party evaluators embedded inside AI labs, coordinated safety standards among companies in democratic countries, and eventually broader international coordination, potentially involving government, partly to address antitrust concerns that could otherwise block companies from legally coordinating on safety limits together. The timing isn't incidental. This follows the resignation of AI researcher Jacob Coxon, who left OpenAI for Anthropic before leaving the industry entirely, accusing both companies of "gambling with our lives." Congress responded fast: Senator Bernie Sanders cited the Amodei-Musk-Altman alignment directly in renewing calls for federal AI oversight. Worth including the skepticism this has drawn, not just the alignment. David Sacks has characterized the coordinated pacing proposal as, in effect, the industry's biggest players building a cartel, restricting competition under a safety label. That's a legitimate tension: proposals for companies to jointly agree on pacing and standards can look identical whether the motive is genuine risk mitigation or reduced competitive pressure, and it's often difficult to distinguish the two from the outside. There's also a practical question closer to home: the Trump administration has favored a deregulatory approach to AI so far, so it's unclear how much appetite exists in Washington for the kind of government role Amodei is describing.