The federal agency responsible for defending America’s critical infrastructure from cyberattacks has lost roughly one-third of its workforce since early 2025. And it’s about to get worse.
The Cybersecurity and Infrastructure Security Agency, better known as CISA, is facing a proposed $707 million budget cut in Fiscal Year 2027 that could eliminate another 766 full-time positions. This is happening at the exact moment AI-driven cyber threats are evolving faster than most organizations can patch their systems.
The AI threat that changed the calculus
On April 7, 2026, Anthropic released Claude Mythos Preview, an AI model capable of identifying thousands of zero-day vulnerabilities and executing autonomous attacks.
Zero-day vulnerabilities are security flaws that software makers don’t yet know about. They’re the most dangerous kind of exploit because there’s no patch available when attackers strike. Historically, finding them required elite hackers with deep expertise spending weeks or months probing systems. Claude Mythos Preview compressed that timeline dramatically.
