Google just paid $32 billion for a six-year-old security startup, and the Valley is finally saying the quiet part loud about what AI infrastructure actually costs.

The Signal

Wiz, founded in 2020, became the largest venture-backed acquisition in history at $32 billion. Google tried to buy them in 2024, got turned down, and came back with more money after sweating through antitrust reviews on both continents. That should tell you something about desperation.

Index Ventures calls this the "Deal of the Decade" because Wiz sits at the convergence of AI, cloud infrastructure, and security spending. Translation: as companies race to deploy AI agents, they're discovering their security posture is swiss cheese. Every agent that touches your data, every API call to an external model, every autonomous decision is a new attack surface. The legacy security stack was built for humans clicking buttons, not for thousands of agents making microsecond decisions across distributed infrastructure.

Google isn't buying Wiz for their current revenue. They're buying insurance. As the agent economy scales, whoever controls cloud security controls the chokepoint. Amazon has GuardDuty. Microsoft has Sentinel and Defender. Google just realized they can't build fast enough to compete, so they bought the company that already guards multi-cloud environments. This is infrastructure consolidation disguised as an acquisition.

The price tag also reveals something else: the market now prices security tools at AI-era multiples. Six years from zero to $32 billion means investors believe the agent economy creates existential security risks worth paying almost anything to solve.

The Implication

If you're building in the agent economy, security isn't a feature you bolt on later. It's becoming the most valuable part of the stack. Watch for more mega-acquisitions in identity management, data governance, and agent behavior monitoring. The companies that let AI systems move fast without breaking everything are the new infrastructure layer, and they're going to get paid like it.

Source: TechCrunch AI