AI security company HiddenLayer announced a $100 million Series B on September 2, led by Delta-v Capital with participation from Ten Eleven Ventures, Morgan Stanley, M12 (Microsoft's venture fund) and Booz Allen Ventures. The round brings HiddenLayer's total funding to $156 million and comes as the company says annual recurring revenue has grown more than tenfold over the past year.

What HiddenLayer does

HiddenLayer builds security tooling for agentic, generative and predictive AI systems, positioning itself as a layer that monitors and defends AI applications the way traditional endpoint or network security tools defend conventional infrastructure. Its existing Agentic Runtime Security module watches AI agents while they operate in production, looking for behavior like prompt injection, unauthorized tool use, or an agent drifting outside its intended scope. With this round, the company is launching Agent Harness Security, a new product that extends that runtime protection specifically to autonomous AI coding agents — the class of tools, from Claude Code to Cursor to OpenAI's Codex, that enterprises are increasingly plugging directly into their development environments with the ability to write and execute code.

Why now

HiddenLayer says the new capital will go toward expanding its enterprise platform, with runtime and agent-specific protection as the near-term focus. Secondary-market pricing referenced by Yahoo Finance puts HiddenLayer's post-money valuation at roughly $387 million. The round follows a wave of similarly-themed raises in the past two months, including AI Score's $5.4 million seed for agent governance and visibility, and AIR's $50 million seed for vetting the plugins and skills agents use — all pointing to investors treating agent security as a distinct, fast-growing category rather than a feature bolted onto existing AI-safety products.

Why it matters

Coding agents that can read, write and execute code inside a company's actual repositories and infrastructure carry a materially different risk profile than a chatbot answering questions: a compromised or manipulated coding agent can introduce vulnerabilities, exfiltrate credentials, or make production changes with far less human review than traditional software delivery involves. HiddenLayer's bet with Agent Harness Security is that this specific slice of the market — autonomous agents with write access to code and systems — will need dedicated runtime controls distinct from the input/output filtering that generative-AI security tools have focused on to date. The size and investor list of the round, including strategic backing from Microsoft's venture arm and Morgan Stanley, is a signal that agent security is being treated as core enterprise infrastructure rather than a niche compliance purchase.