HiddenLayer Secures $100M as AI Agent Security Becomes Enterprise Priority

TL;DR

• HiddenLayer has raised a $100M Series B funding round to expand its AI security platform, reflecting surging enterprise demand for protection as generative AI and autonomous agents move into production.
• The core challenge for enterprises is no longer just securing AI models, but monitoring and governing the entire AI stack — including agents, tools, plugins, and add-ons that create new and often invisible attack surfaces.
• With AI security spending projected to grow into a multi-billion dollar market, investors and competitors are racing to build next-generation solutions for threat detection, model integrity, and agent behavior monitoring.

From Model Security to Agent Security

HiddenLayer, the Austin-based AI security startup, has secured $100 million in new funding as enterprises confront a harsh reality: securing artificial intelligence is no longer just about the model. The round, one of the largest to date in the AI security space, signals a major shift in enterprise priorities from AI experimentation to AI protection at scale.

The funding arrives at a moment when AI adoption has exploded beyond chatbots and copilots. Companies are now deploying autonomous AI agents that can reason, take actions, use tools, access sensitive data, and interact with other systems. That autonomy is powerful, but it has also created an entirely new class of security risks that traditional cybersecurity tools were never designed to handle.

Why the AI Stack Is Now the Attack Surface

For the past two years, most AI security conversations focused on prompt injection, data poisoning, and model theft. Those threats remain critical, but security leaders say the attack surface has widened dramatically.

Modern enterprise AI deployments are rarely a single model. They are complex stacks composed of large language models, vector databases, retrieval-augmented generation pipelines, third-party tools, browser plugins, and fleets of AI agents with persistent memory and system-level permissions. Each component is a potential entry point.

An agent that can read email, execute code, or make financial transactions can be hijacked through a malicious prompt hidden in a document, a compromised tool, or a poisoned data source. Security researchers have already demonstrated attacks where agents were tricked into exfiltrating data, bypassing approval workflows, or executing unauthorized actions — all while appearing to function normally.

This is why monitoring has become the central challenge. Enterprises need real-time visibility into what their agents are doing, what tools they are calling, what data they are accessing, and whether their behavior has drifted from its intended purpose.

Inside HiddenLayer's Platform Play

Founded in 2022, HiddenLayer initially made its name with protections for machine learning models, including its AISec Platform designed to detect and block adversarial attacks on AI systems. The company has since expanded its focus to cover the broader AI lifecycle, including detection and response for generative AI and autonomous agents.

The new capital will be used to accelerate product development around AI agent monitoring, runtime protection, and governance, as well as to expand go-to-market operations as demand from Fortune 500 customers grows. The company says its platform provides visibility into model inputs and outputs, detects anomalous agent behavior, and can automatically block malicious activity before it propagates through enterprise systems.

Investors are betting that this type of purpose-built AI security layer will become as fundamental as endpoint or cloud security. The round was led by a mix of cybersecurity-focused venture firms and strategic investors, underscoring confidence that AI security is moving from a niche concern to a core enterprise budget item.

A Market Racing to Keep Up

HiddenLayer is not alone in chasing the opportunity. The explosive growth of enterprise AI has triggered a wave of activity across the security industry. Startups specializing in AI firewalling, prompt security, AI posture management, and agent observability have all raised significant capital in the last 12 months. At the same time, large cybersecurity incumbents are rapidly adding AI security modules to their existing platforms through product launches and acquisitions.

Analysts estimate the AI security market will grow from under $2 billion in 2024 to more than $10 billion by 2028, driven by regulatory pressure, high-profile AI-related incidents, and the sheer scale of enterprise deployment. Frameworks from NIST and new governance requirements in the U.S. and EU are also pushing companies to prove their AI systems are secure, auditable, and compliant.

What Comes Next for Enterprise AI Security

The HiddenLayer raise makes one thing clear: enterprises are no longer asking if they need to secure their AI — they are asking how.

As organizations move from dozens of AI pilots to hundreds of agents in production, security teams will need to shift from reactive model scanning to continuous, real-time protection of the entire AI ecosystem. That means monitoring not just what goes into a model, but what an agent does after it reasons, what tools it trusts, and what add-ons it connects to.

For security vendors, the race is on to define the standard for that protection. For enterprises, the message is urgent: every new agent deployed without visibility and control is a new risk waiting to be exploited.

AndroGuider Team

Articles written by the AndroGuider team. We try to make them thorough and informational while being easy to read.