AI Weekly Issue #495: Musk, Zuckerberg killed Trump's AI safety order in three phone calls
Summary
Over a single weekend, significant shifts occurred in AI governance, security, and industry economics. Elon Musk, Mark Zuckerberg, and David Sacks reportedly killed a draft AI safety executive order from the Trump administration in three phone calls on May 25th, 2026, preventing public comment. Concurrently, Anthropic closed a $30B+ funding round, nearing its first operating profit with Claude Code generating \$1B ARR, while Microsoft cancelled its internal Claude Code pilot after token billing consumed its entire annual AI budget. A novel cross-registry supply chain attack, TrapDoor, compromised 34 packages across npm, PyPI, and Crates.io, leveraging AI config files. CISA also reported 15,000 attacks exploiting a Drupal SQL flaw. The White House approved \$9B for NSA Blackwell chips and overruled the Pentagon to maintain Anthropic's presence in classified networks, establishing a "no Americans' data" contract standard.
Key takeaway
For AI auditors, expand security pipelines to include AI agent config files like ".cursorrules" and "CLAUDE.md", now proven attack vectors. If you are planning federal AI compliance, shift focus from executive orders to state laws and the EU AI Act. AI buyers rolling out coding tools must model token spend rigorously; usage-based billing can quickly exhaust budgets. Recruiters hiring from US universities should assume AI-assisted coursework, making interview design crucial for accurate assessment.
Key insights
Federal AI governance is centralizing, while AI costs and security risks demand new operational vigilance.
Principles
- AI governance is shifting to direct West Wing control.
- Usage-based AI billing reveals true enterprise costs.
- AI config files are new software supply chain attack vectors.
In practice
- Add AI agent config files to code review pipelines.
- Model per-developer token caps into AI contracts.
- Design interviews to assess AI-assisted coursework.
Topics
- AI Governance
- Supply Chain Attacks
- AI Procurement
- Large Language Models
- Token Economics
- AI Safety Policy
Best for: CTO, VP of Engineering/Data, Investor, Director of AI/ML, Consultant, Policy Maker
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by AI Weekly — AI News & Updates.