Now Playing
AI Pre-Deployment Rules, $218B Surge & China's $98B Bet
Audio is available on Spreaker — see link below.
The White House is studying an FDA-style executive order requiring frontier AI models to pass federal safety reviews before release — while corporate AI investment hits $218 billion and China commits up to $98 billion in AI capex for 2025. Can NIST's underfunded CAISI keep pace with any of it?
Audio is available on Spreaker — see link below.
The White House is now actively studying an executive order that would require frontier AI models to pass government safety reviews before they can be released to the public. That's a significant shift.
The review process the White House is considering would run through NIST's Center for AI Standards and Innovation, known as CAISI. Originally established under Biden as the AI Safety Institute, it was rebranded under the current administration and is now the primary government contact point for frontier model testing.
While the regulatory picture develops, the investment side is accelerating in a way that makes the capacity question even more pressing. Corporate AI spending jumped twenty-two percent to two hundred eighteen billion dollars in twenty twenty-four.
China is not sitting this out. The country is projecting eighty-four to ninety-eight billion dollars in AI capital expenditure for twenty twenty-five, with fifty-six billion of that coming from direct government investment.
There's one variable that cuts across all of this: power. Global data center electricity consumption is projected to reach nine hundred forty-five terawatt hours by twenty thirty.
The signal to track here is sequencing. The executive order, if it moves forward, arrives into a regulatory environment where the enforcement agency is underfunded, the investment cycle is moving faster than any review process could currently handle, and the competitive pressure from China makes slowdown politically costly.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.