Why AI Agent Security Is Becoming Its Own Category

Mave Security's $34 million combined raise, with annual recurring revenue growing more than 400% year over year, shows agent security has moved from a feature pitch to a standalone funding category.

Trace Cohen
By Trace Cohen · Value Add Pulse AI Desk
Founder & Editor-in-chief, Value Add VC
2 min read
Why AI Agent Security Is Becoming Its Own Category: key figures
ShareXLinkedInEmail

THE RUNDOWN

1

Mave Security's combined $34 million -- a $28 million Series A led by Red Dot Capital Partners plus an earlier $6 million seed -- came with ARR growing more than 400% year over year, a growth rate that justifies the fast follow-on.

2

The company connects a customer's existing security tools into one agentic response system, rather than replacing them -- a lower-friction sale than asking enterprises to rip and replace.

3

A roughly 20-person team doubling in size signals the company is scaling go-to-market, not just engineering, which is the harder part of enterprise security sales to get right.

4

The risk for any agent-security startup is that the underlying agent platforms (OpenAI, Anthropic, Google) could build equivalent governance natively, compressing the standalone market.

The VC Read

Value Add VC analysis

The diligence question for agent-security startups is multi-platform coverage: does the product work identically across OpenAI, Anthropic and open-weight agent stacks, or does it quietly work best on just one? A 400% ARR growth rate built on a single-platform integration is a much weaker moat than the same growth rate built on genuine cross-platform governance.

Analysis

Mave Security's $34 million combined raise is a useful marker for how fast AI agent security has moved from a feature inside broader security platforms to a funding category with its own round structure and growth benchmarks. The total breaks down into two separate rounds:

  • Series A -- $28 million, led by Red Dot Capital Partners.
  • Seed -- $6 million, backed by Dell Technologies Capital and Emerge.

“- Series A -- $28 million, led by Red Dot Capital Partners.”

The company's pitch is integration, not replacement: it connects a company's existing security tools into one agentic response system rather than asking enterprises to swap out what they already run. That lower-friction sale appears to be working -- annual recurring revenue grew more than 400% year over year, and the roughly 20-person team is doubling, a sign the company is investing in sales and support capacity to match demand rather than just engineering headcount.

The category question is whether this growth rate is Mave-specific or a signal about the whole agent-security space. As more enterprises put AI agents into production with real tool access and data permissions, the governance layer watching what those agents actually do becomes a budget line item in its own right, not a checkbox inside a broader security suite -- which is exactly the bet every investor writing agent-security checks right now is making. A 400%-plus ARR growth rate, even off a small base, is the kind of number that turns a seed-stage security feature into a category other founders start building competitors to chase.

The risk is platform risk: OpenAI, Anthropic and Google all have obvious incentives to build governance and monitoring directly into their own agent platforms, which would compress the addressable market for any standalone agent-security vendor that doesn't differentiate clearly on breadth across multiple model providers. For comparison, Watchtower Labs' $35 million seed for military AI, also raised recently, shows check sizes in adjacent AI-security verticals running similarly large right now, even though the two companies sell into very different buyers -- one into enterprise security teams, the other into defense procurement.

ShareXLinkedInEmail

Key Sources

1 source

THE WIRE in your inbox— Tech, startup & VC news with The VC Read, a few times a week. Free to subscribe, no spam.

No spam, unsubscribe anytime. Sent via Substack · Privacy · or subscribe on Substack ↗