Last Tuesday, in a room inside the White House, six men stood beside President Donald Trump and signed a document he called "morally binding." Elon Musk, Mark Zuckerberg, Jensen Huang and Dario Amodei were among them, agreeing to a loose framework for regulating artificial intelligence that relies, in Trump's own words, on "tremendous self-policing." The signing looked ceremonial. What happened beforehand was not.

According to Semafor, Zuckerberg did not simply attend the gathering — he helped write the document that emerged from it. The idea reportedly began as a conversation between the Meta chief executive and House Speaker Mike Johnson at a state dinner honoring China's Xi Jinping. Zuckerberg then circulated a draft before the CEOs convened, and Huang backed it. The result is a strange hybrid: a private industry memo that now carries the authority of the Oval Office.

Few governments in Washington, London, or Brussels have ever let the companies they regulate draft the regulation itself quite so openly. The symbolism deepened days later, when Trump ordered that federal documents replace the phrase "artificial intelligence" with "super intelligence." Musk complied almost instantly, renaming a SpaceX subsidiary from SpaceXAI to SpaceXSI — a small gesture that signaled something larger: his return to government influence, over a year after his turbulent stint leading the Department of Government Efficiency. That same week, Defense Secretary Pete Hegseth appointed Musk, alongside Palmer Luckey and Emil Michael, to a new taskforce on the future of warfare, called Project Meridian.

For allies and rivals abroad, especially in Europe, where most AI infrastructure is still imported from American firms, the implications are hard to ignore.