HiddenLayer AI Security Platform

HiddenLayer, Inc. is an AI security company based in Austin, Texas, founded by security researchers who uncovered the first AI-specific malware and led by CEO and co-founder Chris Sestito. It protects agentic, generative and predictive AI applications for enterprises and governments, addressing shadow AI, malicious third-party models, prompt attacks, model theft and unsafe agent behaviour.

The platform has four modules. AI Discovery builds a living inventory of models, applications and AI assets, including shadow AI. AI Supply Chain Security scans models for malware, backdoored weights, vulnerable dependencies and unknown components before deployment. AI Attack Simulation runs continuous red teaming against agentic and generative applications. AI Runtime Security acts as a firewall that detects and responds to prompt injection, data leakage, model extraction and unauthorised tool use in production, and Agent Harness Security extends it to AI coding agents. The platform is model-agnostic and agentless and needs no training data.

• AI Discovery: Builds a living inventory of models, AI applications and assets, including shadow AI.
• Model scanning: Detects malware, backdoored weights, vulnerable dependencies and unknown components before deployment.
• AI Attack Simulation: Runs continuous adversarial testing against agentic and generative applications.
• AI Runtime Security: Detects and responds to prompt injection, data leakage, model extraction and unsafe tool use in production.
• Agent Harness Security: Gives runtime visibility, threat detection and inline enforcement for AI coding agents.
• Governance and posture: Classifies risk and applies consistent policy across every model in the enterprise.

A Vietnamese bank can scan open-source models downloaded for credit scoring or chatbots for malware and backdoors before deployment. An insurer can put runtime protection in front of its claims assistant to block prompt injection and data leakage. A technology company adopting coding agents can use Agent Harness Security to monitor and block unsafe actions.

Enterprises can validate third-party and open-source models before they reach production, reducing the risk of backdoors and malicious code in the AI supply chain. Runtime protection blocks prompt attacks, model extraction and data leakage without exposing weights, prompts or customer data to the vendor. Consistent governance and risk classification across every model help meet regulatory requirements. Native connectors for cloud, CI/CD, SIEM/SOAR, API gateways and MLOps tools fit into existing security operations.

Most companies build on models from outside sources such as open-source hubs, which brings unknown code into production, while autonomous agents now write and ship code. Security for AI is shifting from point tools to lifecycle platforms covering inventory, model scanning, red teaming and runtime defence. Banks, insurers and public agencies in Vietnam and Southeast Asia that fine-tune open models for fraud detection or customer service need this supply chain assurance.

HiddenLayer, Inc. is based at 14900 Avery Ranch Blvd, Austin, Texas. On 2 September 2026 it announced a 100 million USD Series B led by Delta-v Capital with Ten Eleven Ventures, Morgan Stanley, M12 and Booz Allen Ventures, reporting more than 10x ARR growth and over 50 new platform customers. It is available on AWS, Microsoft Azure and Google Cloud marketplaces and partners with AWS and Databricks.