For years, enterprise security has been a game of perimeter defense and identity management. But as companies move from simple chatbots to autonomous AI agents – software that can execute tasks, access data, and make decisions on behalf of a user – the old playbook is fraying. The recent $100M Series B funding for Austin-based HiddenLayer, announced on September 2, 2026, is the clearest signal yet that the industry is moving to formalize a new, standalone category: AI agent security.
The round, led by Delta-v Capital with participation from Ten Eleven Ventures, Morgan Stanley, M12 (Microsoft’s venture fund), and Booz Allen Ventures, is not just a capital injection. It is a validation of a shift in how enterprises view their software stacks. When investors like M12 and Booz Allen put significant weight behind a startup, they are signaling that the risks associated with agentic AI are no longer theoretical concerns for the R&D lab; they are immediate operational liabilities for the enterprise.
The momentum is undeniable. In just the last five weeks, the market has seen over $150 million in capital flow into this specific niche, split between HiddenLayer’s $100 million round and a $50 million raise by AIR. This is not a slow trickle of interest; it is a rapid consolidation of resources around the idea that agents require a dedicated layer of protection.
As HiddenLayer CEO Chris Sestito noted, the funding is intended to help the company “keep growing the purpose-built team and platform required to meet that moment as agentic AI becomes core to how enterprises operate.” That “moment” is the transition from AI as a passive tool to AI as an active participant in business workflows.
This shift has birthed two distinct technical requirements: Agentic Runtime Security and Agent Harness Security. These are not just buzzwords; they represent the need to monitor agents while they are actively working – ensuring they don’t deviate from their intended logic – and to secure the “harness” or the environment in which these agents operate. Mark Hatfield of Ten Eleven Ventures put it bluntly: “AI needs its own category of protection.”
The broader ecosystem is already reacting. Broadcom has introduced AgentMinder at VMware Explore, and Okta is pushing forward with Agent SSO. These moves suggest that the industry is moving toward a standard where agents are treated as first-class citizens in the enterprise, requiring their own authentication, monitoring, and guardrails.
Austin’s emergence as a hub for this category is no accident. The city already hosts CrowdStrike, which relocated its headquarters there in 2021, and SailPoint, the identity governance pioneer founded in Austin. The University of Texas at Austin’s computer science program has been a steady pipeline for AI and security startups, and HiddenLayer’s growth adds another anchor to a cluster that is increasingly defined by the intersection of security and autonomous systems. When a $100 million round lands in a city that already houses two of the largest identity and endpoint security companies in the world, the ecosystem effect compounds.
HiddenLayer plans to use its new capital to deepen its enterprise platform, expand these specific security offerings, and build out channel relationships while pushing into international markets. This is the classic playbook for a company trying to define a category before the market becomes crowded with generalist security vendors.
However, there is a practical caveat to this rapid category formation. While the investment signals a clear demand for security, the actual implementation of these tools remains a work in progress. Enterprises are still figuring out how to balance the autonomy of AI agents with the rigid control required by security teams. Buying a security platform is only the first step; the real challenge will be integrating these tools without stifling the very productivity gains that agents are meant to provide.
The emergence of a standalone category for agent security is a logical response to the complexity of autonomous systems. If agents are going to be the new workforce, they need a new kind of management. The $150 million spent in the last few weeks suggests that the market has decided it is better to build that infrastructure now than to deal with the consequences of an unsecured agent later.