HiddenLayer, an AI security startup, closed a $100 million funding round as enterprises accelerate efforts to lock down their artificial intelligence deployments against growing threats.
The company addresses a gap in enterprise AI security that has widened as organizations race to deploy large language models and AI agents across their operations. While traditional security tools focus on perimeter defense, HiddenLayer targets the specific vulnerabilities embedded in AI systems themselves. The startup's platform monitors not only AI agents but also the tools, plugins, and add-ons those agents integrate with. This matters because AI agents increasingly operate autonomously, pulling in third-party APIs and extensions that create new attack surfaces.
The funding reflects investor confidence in a security category that barely existed two years ago. As enterprises deploy AI into production, security teams struggle to protect systems that operate differently from legacy software. An AI agent might pull data from a company database, feed it to an LLM, run analysis, and execute actions across multiple tools. Each connection point represents a potential compromise. Hackers can poison training data, inject malicious prompts, or hijack agent behavior through compromised tools.
HiddenLayer competes in a crowded but nascent market. Other players like Robust Intelligence, Lakera, and Promptly Security target similar problems. But HiddenLayer's approach of monitoring the entire agent supply chain. including third-party integrations, positions it differently. The company watches for unauthorized model changes, detects prompt injection attacks, and flags suspicious tool usage patterns in real time.
The $100 million round signals that investors believe AI security will command significant spending. Enterprise customers already grapple with compliance mandates around AI governance. Regulators in the EU and US are beginning to require documentation of AI system safety. Security tools that provide visibility and audit trails become table stakes.
Timing favors HiddenLayer. Organizations running AI agents in customer service, finance, and operations need assurance that agents won't leak sensitive data or execute unauthorized transactions. A compromised agent managing customer support tickets could expose personal information at scale. One infiltrated financial analysis tool could generate fraudulent reports.
The company plans to invest the new capital into expanding its platform capabilities, hiring security researchers, and building detection models trained on real-world AI threats. HiddenLayer also aims to deepen integrations with major cloud providers and AI platforms, ensuring its monitoring tools sit directly in deployment pipelines.
This funding arrives as major tech companies release increasingly powerful AI agents. OpenAI's tool use capabilities, Anthropic's Claude agents, and Agentic frameworks from major cloud providers all push enterprises toward production AI deployments. Each deployment multiplies the attack surface HiddenLayer targets.
The security market has historically followed deployment waves. Each new technology layer, from servers to containers to serverless functions, spawned dedicated security companies. AI agents represent the latest such layer. HiddenLayer enters the market well before mainstream AI adoption peaks, positioning it to capture customers as AI deployments scale across enterprise infrastructure.