Neo matters because investors are funding a control layer for agentic software across the whole enterprise, not just one more detection dashboard.

What Neo Announced

On July 20, 2026, Neo announced a $100 million funding round from Andreessen Horowitz and Bessemer Venture Partners, with Craft Ventures and Merlin Ventures also participating. Neo says it gives SecOps teams software inventory, posture intelligence, attribution, and policy control across AI agents, AI-enabled apps, browsers, identities, and traditional software that is gaining agentic capabilities.

Neo's framing is useful because it does not isolate risk to one agent product. It treats the wider software estate as the problem surface, especially once approved tools start reasoning, invoking tools, and moving through workflows with valid user permissions.

Why This Investment Signal Is Strong

Many security products still assume agents are rare, explicit, and centrally deployed. Neo is betting on the opposite: agentic behavior will show up inside the software companies already run, which means control has to live at the software layer itself.

That is a stronger zero-human signal than another model-scanning tool. It assumes autonomous behavior is becoming normal enough that enterprises need real-time policy enforcement, attribution, and capability mapping before risky activity happens.

Why The Unit Of Control Is Shifting

Traditional governance asks whether a user is allowed to access a system. Agentic governance has to ask what a software entity can actually do once it is inside, which tools it can chain together, and how those behaviors change as vendors ship new features.

Neo is effectively treating “agentic software” as a new enterprise asset class. That is the right lens for zero-human companies, where the software layer itself becomes a moving workforce.

The Take

Neo is a meaningful investment signal because it prices enterprise-wide control over agentic software as a category, not a niche security add-on.

The more autonomy spreads into existing software, the more valuable real-time inventory, attribution, and policy enforcement become as operating infrastructure.

Related: See our earlier research on WitnessAI, NewCore, Runta, and AegisAI.