Musk, Zuckerberg, Sacks kill AI safety order; Anthropic hits $30B, NSA gets $9B chips
In a single week, the White House killed a draft AI safety executive order with industry pressure, then allocated $9 billion in chips to the NSA—pouring fuel on the same fire they claimed needed no oversight.
Elon Musk, Mark Zuckerberg, and David Sacks killed Trump's draft AI safety executive order in three Wednesday-night calls, framing it as 'doomer regulation.' The same Saturday, Anthropic closed a $30B+ round with $10.9B annualized revenue and Claude Code at $1B ARR. Microsoft canceled its internal Claude Code pilot after token billing consumed its entire annual AI budget. CISA flagged 15,000 exploit attempts against a Drupal SQL flaw within 48 hours of patch. The White House overrode the Pentagon to keep Claude inside the NSA while clearing $9B for NSA Blackwell chips.
- Anthropic's $10.9B revenue claim likely stems from a single large contract annualized, not organic growth; treat it as aspirational until audited.
- The $9B NSA chip allocation decouples government AI spending from safety rhetoric, prioritizing capability above all else.
- Microsoft's cancellation of Claude Code due to token billing highlights the real cost of premium AI agents, limiting enterprise adoption beyond early pilots.
Why It Matters
The AI safety debate is now a high-stakes tug-of-war between industry influence and national security priorities.