HiddenLayer raises $100M to defend AI agents from cyber threats

HiddenLayer, a specialist in AI and machine learning security, has closed a $100 million Series B funding round led by Thrive Capital, with participation from existing investors including Ten Eleven Ventures and Constellation Technology Ventures. The Austin-based startup, which emerged from stealth in late 2023 with a platform designed to detect adversarial attacks and data poisoning in AI models, announced the round on April 16, 2025. The financing brings HiddenLayer’s total funding to $145 million and values the company at over $500 million. This rapid valuation surge reflects mounting enterprise anxiety over the security of AI agents—software entities that autonomously perform tasks using tools, APIs, and third-party models.

The company’s core offering, AI Guardian, operates as a runtime security layer that continuously monitors AI agents in production, including their tool use, data inputs, and decision pathways. In a market where most security vendors still focus on traditional endpoints or network traffic, HiddenLayer’s approach targets the unique risks posed by agentic AI: prompt injection, model theft, and supply-chain compromise of AI components. CEO Chris Sestito, a former NSA analyst and Palantir executive, emphasized that the funding will accelerate product development, particularly in supporting multi-agent orchestration platforms and deep integration with observability tools like LangSmith and Weights & Biases.

Recent high-profile breaches have exposed vulnerabilities in AI supply chains. In March 2025, a compromised third-party fine-tuning API led to the leakage of proprietary financial data from a major U.S. bank’s internal AI assistant. HiddenLayer’s platform detected the anomaly within minutes, according to a case study involving JPMorgan Chase’s internal agent, which operates on a multi-cloud architecture for maximum reliability and global reach in financial market monitoring. Such incidents have pushed CISOs to prioritize agent security, with Gartner predicting that by 2026, 70% of enterprises will actively monitor AI agents using specialized runtime protection tools, up from less than 5% in 2024.

The funding round arrives amid a broader consolidation in the AI security sector. Competitors like Protect AI, Calypso AI, and Vanta are racing to expand beyond vulnerability scanning into runtime monitoring. Protect AI, for instance, recently acquired AI security startup Lasso Security for $200 million, signaling a land grab for agent-level controls. HiddenLayer’s differentiation lies in its deep instrumentation of model internals and real-time behavioral analysis, rather than relying solely on API logs or external telemetry. Industry insiders note that this technical depth resonates with regulated sectors such as finance and healthcare, where model lineage and auditability are non-negotiable.

The infusion of capital also reflects a strategic pivot toward global markets. With operations in the U.S. and the U.K., HiddenLayer plans to expand into Asia, targeting financial institutions in Singapore and Tokyo that are rapidly deploying AI agents for trading, fraud detection, and regulatory compliance. The company’s platform has already been adopted by early adopters like Capital One and a Tier 1 U.S. healthcare provider, where it monitors AI-driven claims processing agents. Analysts at McKinsey estimate that the AI runtime security market could reach $3.2 billion by 2028, growing at a compound annual rate of 78%.

This surge in enterprise demand is part of a larger tectonic shift in the Quantum & Computing landscape. AI agents are increasingly powered by quantum-inspired optimization algorithms and run on hybrid cloud architectures that blend classical and quantum processing units. While quantum computing remains years away from widespread deployment, the integration of quantum machine learning models into agent workflows has already begun in sectors like logistics and drug discovery. HiddenLayer’s investment signals a convergence: securing AI agents today is not just about preventing data leaks, but ensuring that tomorrow’s quantum-accelerated agents operate within trusted, auditable boundaries. Competitors like Qrypt and Cambridge Quantum have started integrating quantum-resistant cryptography into their security stacks, anticipating a future where AI agents interact with quantum data sources.

The geopolitical dimension adds further urgency. National security agencies in the U.S. and EU are scrutinizing AI agent ecosystems for vulnerabilities that could be exploited by state actors. In February 2025, the U.S. Cybersecurity and Infrastructure Security Agency (CISA) issued a warning about adversarial manipulation of AI agents in critical infrastructure, urging adoption of runtime monitoring tools. HiddenLayer’s platform has been cited in closed-door briefings as a reference architecture for agent defense, suggesting alignment with government priorities. Meanwhile, in China, companies like Huawei and Baidu are developing competing AI security frameworks, creating a bifurcated landscape where Western and Eastern approaches to agent security may diverge.

Looking ahead, the next phase of AI security will likely revolve around standardization and interoperability. The Open Worldwide Application Security Project (OWASP) is expected to release a Top 10 list for AI agents in late 2025, mirroring its famous web security list. HiddenLayer has already contributed to the initiative, and its leadership expects certification requirements to emerge within two years. Enterprises should prepare for increased scrutiny from boards and regulators, particularly around agent autonomy and third-party tool usage. The $100 million infusion positions HiddenLayer as a frontrunner in this high-stakes race, but the real test will be whether its technology can scale across diverse agent ecosystems without introducing latency or operational friction. For the industry, the message is clear: securing AI is no longer optional—it’s existential.