AI broke the patch. This $60m startup is betting on not patching.
What happened
Act Security, an Israeli startup founded by the team behind Medigate, stepped out of stealth mode with a $60 million funding round. The company launched in 2025 with a clear challenge to conventional cybersecurity wisdom: patching software vulnerabilities is no longer a reliable path to safety. Instead, they propose a new approach powered by AI agents and cloud-native architecture that sidesteps traditional patching.
Why it matters
For decades, cybersecurity has fixated on patch management as a core defense. Act Security’s argument flips that mindset, highlighting patching as slow, incomplete, and increasingly ineffective against advanced threats. The startup’s approach targets how organizations control cloud access and manage risk dynamically, relying on AI agents to detect and respond to threats without waiting for patches to roll out.
This pressures legacy security models that assume patches are the primary fix for vulnerabilities. For operators, it means shifting resources away from constant patch chasing toward deploying smarter AI-driven controls that reduce attack surfaces in real time. For investors, the $60 million launch signals confidence in security strategies that break from the patch-first orthodoxy, betting AI can automate risk reduction more effectively.
What to watch next
Track how Act Security’s AI-driven cloud access model performs in practice against established patch-based solutions. Watch if other startups or security incumbents start abandoning patch-centric models, especially in regulated or critical infrastructure sectors where patch windows are tight and risk tolerance is low.
Also observe customer adoption patterns. Enterprises reluctant to overhaul existing patch processes will test how much AI agents can offset risk without adding complexity. Success or failure here will influence broader trust in AI’s ability to reshape operational security from a reactive to a proactive footing.
AI Quick Briefs Editorial Desk
