An essay about slowing down cost Nvidia 3% and SoftBank 13%

Dario Amodei published "We Must Pace the Frontier" on September 12. Altman and Musk agreed within hours. Hassabis called the direction correct. By Monday morning the market had priced it.

Nasdaq 100 futures fell 1.5% and S&P 500 futures 0.8%. Nvidia dropped 3%, AMD 5.7%, and Arm, Marvell and CoreWeave each about 6%. In Asia, SoftBank fell 13%, SK Hynix 5.3%, Samsung 2.8%, Kioxia over 5%, with the KOSPI down 2.5%. Altman's separate statement that an IPO "right now would be an ill-advised moment to go public" compounded it.

The distribution of damage is the detail nobody's talking about. The labs didn't take the hit. Memory and capex did. Hynix, Samsung, Kioxia, Micron-adjacent names, the GPU vendors and the neoclouds. The market read "the labs might slow down" and sold the people who sell them hardware, which is a coherent read of a supply chain where the buyers just floated demand destruction as a policy goal.

Politically the proposal is dying faster than it was priced. Trump told AP he sees "a lot of negative forces that are... bringing up things that won't happen" and framed it as a race. Speaker Mike Johnson said Congress "has to resist... some sort of emergency moratorium." White House AI adviser David Sacks endorsed the substance and rejected the mechanism, telling the labs to stop pretending antitrust law has to be suspended "so you can form a cartel," and questioning whether METR is independent enough to referee given its entanglement with Anthropic's investors and staff. His closing line: "the easiest way not to build superintelligence is for you to agree not to build it."

China's foreign ministry called it "fear-mongering" and Global Times called it a Cold War playbook, while Xi used the BRICS summit in New Delhi to propose a BRICS open-source AI community. Washington and Beijing landed on the same side against the labs inside 48 hours.

And the labs don't agree with each other either. CNN reports Anthropic, OpenAI and Google have been in working-group meetings since July on a private standards body and hold incompatible models of it. Amodei wants an FAA that can block a release on day one. Hassabis wants a FINRA, industry-funded with federal oversight and voluntary pre-release review. Altman reportedly thinks the labs will have to build it themselves without government backing. Cohere's Aidan Gomez broke publicly with all of them, comparing lab-written safety standards to bond rating agencies and proposing publicly funded international standards instead.

Amodei himself conceded the day after publishing that China is the unsolved part of his own plan.

So: a proposal with no agreed enforcement mechanism, rejected by the administration and Beijing, opposed by at least one frontier CEO, and already priced into memory stocks. The essay moved capital before it moved policy, and it may never move policy at all.