HiddenLayer, an Austin AI security firm, announced a $100 million Series B round on Wednesday. The capital will be used to develop agentic runtime security tools that protect AI coding agents. The funding will expand its platform, integrate with development pipelines, and address growing threats to generative AI applications.
HiddenLayer's Series B was led by existing backers and new venture partners, though the exact names were not disclosed. The capital will accelerate development of agentic runtime security, a capability that monitors AI code execution in real time. Company leaders say the surge in generative AI projects has created new attack surfaces, making runtime protection essential for developers and enterprises.
The fresh capital will be funneled into building agentic runtime security, a technology that watches AI code as it runs and blocks malicious behavior before it reaches production. Executives note that as AI models become more autonomous, the risk of code‑level attacks rises, prompting the need for continuous, real‑time safeguards. The investment underscores growing confidence in real‑time AI defense solutions.
If deployed widely, agentic runtime security could let developers embed safety checks directly into AI pipelines, reducing reliance on external audits. The approach promises to lower the incidence of model poisoning and data leakage, two emerging threats identified by security analysts. However, integration complexity and performance overhead may slow adoption across smaller teams.
Analysts expect HiddenLayer's investment to accelerate a new security paradigm where runtime monitoring becomes standard for AI applications. The company aims to launch its agentic tools by early 2027, positioning itself as a key defender against evolving AI‑driven threats. Success could reshape how enterprises secure generative AI workloads and influence future funding trends.
What does agentic runtime security aim to protect? It continuously monitors AI code as it executes, blocking malicious actions before they affect models or data. The technology targets code‑level attacks such as model poisoning and unauthorized data access.
When will the new security tools be available? The company says the first version of its agentic runtime platform is slated for release in early 2027, with broader rollout later that year. Developers can expect a beta phase starting mid‑2026.
How might this funding influence the AI security market? Analysts expect the influx of capital to spur competing startups to develop similar runtime defenses, intensifying competition and potentially lowering costs for enterprises seeking AI protection. This could also drive innovation in related security technologies.