After February 5, 2026, 150,000 people were running autonomous AI agents on personal computers — with access to their files, browsers, and systems. Enterprise companies were deploying agent teams with access to real infrastructure. The attack risk for critical systems — power grids, financial networks, air traffic control — expanded dramatically.
Every AI agent running autonomously — every API call, every agent instance, every workflow — needs edge security, filtering, and protection from attacks. The attack surface didn't just grow. It changed shape. Traditional perimeter security was designed for humans logging in during business hours. Agent traffic is continuous, automated, and operates at machine speed. The security architecture needs to be fundamentally different.
And there's a second problem: governance. When an autonomous agent has access to a company's systems and data, somebody has to ensure it doesn't execute an unverified action. Somebody has to be the air traffic controller — the layer that sits between the agent's decision and the real-world execution. Without that layer, an autonomous bot can go rogue and disrupt a critical system with no human in the loop.
We added positions in both edge security and AI governance as part of our February 2026 repositioning — based on the thesis that autonomous agents would make these capabilities mandatory infrastructure.
Thirteen days later, Operation Epic Fury launched. On February 28, the United States conducted a massive military campaign. Over 1,000 targets were struck in the first 24 hours. The AI governance platform we had just invested in coordinated the operation — in the most high-stakes, real-world environment imaginable.
That was the fastest thesis validation we've experienced in the portfolio. We bought the governance layer because we believed autonomous agents needed air traffic control. Less than two weeks later, the U.S. military proved it — at the scale of a major conflict.
The security side is equally urgent. The edge network that filters every API call and every agent request has seen traffic volumes spike. As more enterprises deploy autonomous agents, the demand for edge security scales linearly with agent adoption. This isn't cyclical. It's structural.
We hold positions in both edge security and AI governance — including one company we've now invested in twice, first riding it from entry to a 6x return, then re-entering at a different moment for different reasons. Members get the full history.
See the Security Positions →Security and governance are the fastest-evolving layer of the AI stack. We track every development — members get the positions and the real-time research.
Subscribe — $297/quarter →