Is the AI industry really ready to slow down?
TechCrunch Anthony Ha ● Covered by 110 sources
Anthropic wants to “pace the frontier”; Nvidia says the backlash is fake. The AI slowdown talk is getting loud, but the actual plan still looks fuzzy.
Based on reporting by TechCrunch, Anthony Ha — 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
AI safety is suddenly the hottest argument in the room, and not everyone is buying the sincerity of the people leading it. Anthropic CEO Dario Amodei has put forward a plan to “pace the frontier,” while Nvidia CEO Jensen Huang has been publicly backing Donald Trump’s line that AI backlash is a hoax and regulation isn’t needed at all.
On TechCrunch’s Equity podcast, Anthony Ha, Kirsten Korosec and Sean O’Kane picked apart whether the new slowdown talk means anything real. Sean said he had been skeptical just a week earlier, and the new wave of support from Amodei, OpenAI CEO Sam Altman and even Elon Musk did make him pause. But he kept coming back to the same complaint: there still isn’t much detail on what “slowing down” actually means, or why these leaders think the risks are as urgent as they say.
That vagueness came up again and again. Kirsten pointed to the broad outline of the proposals: independent third-party evaluators inside companies like Anthropic or OpenAI, major AI firms in democratic countries coordinating on safety standards and limits, and some kind of international coordination. But she also said the idea starts to look a lot less like consensus and a lot more like the beginnings of a cartel.
The most pointed disagreement was over whether the market and existing rules are enough. Sean said that would make more sense in a market with real consumer pressure and a federal government willing to enforce regulations broadly. He doesn’t think that describes AI right now, especially as the money has shifted toward enterprise customers and companies can absorb losses more easily. A business customer isn’t going to ditch OpenAI’s Codex for Anthropic’s Claude Code over a principle, and that’s the problem in a nutshell.
Huang’s role in all this also raised eyebrows. Sean described him as the industry’s “adult in the room,” but Kirsten argued that his appearance with Trump on-stage looked staged and underlined how much Nvidia benefits from an AI boom without brakes. That’s the quiet part of the whole debate: the loudest calls for restraint are arriving at the exact moment the biggest winners still have every reason to keep the pedal down.
My take — AI-written commentary, not fact-checked reporting
The AI crowd loves the word “pace” because it sounds responsible without committing to much. That’s the pattern here: safety talk gets polite fast, then turns into corporate choreography. If the system only works when everyone behaves nicely, it’s not a system, it’s a press release.
Read more about this at: TechCrunch