NewsPublished 16 Sep 2026

AIUC Raises $40 Million to Certify Enterprise AI Agents

Source SecurityWeek

Description

AIUC, a cybersecurity startup, has raised $40 million to advance its AIUC-1 standard, which evaluates and certifies enterprise AI agents against risks like jailbreaks, prompt injections, anomalous behavior, and data leaks. The standard subjects AI models to approximately 5,000 adversarial risk scenarios and includes quarterly audits to address emerging threats. AIUC-1 is already used to certify several popular AI agents. This initiative aims to help enterprises overcome security review bottlenecks for AI adoption by providing evidence of security and reliability.

Impact

The AIUC-1 standard provides enterprises with a structured approach to assess and certify AI agents for security risks, potentially reducing the risk of deploying vulnerable or malicious AI systems. By identifying and mitigating risks such as jailbreaks and prompt injections, the standard helps improve the security posture of AI deployments. However, this is a security assurance and certification initiative rather than a vulnerability or exploit itself, so it does not represent a direct threat but rather a mitigation and risk management advancement.

Defensive guidance

This content describes a security certification and risk assessment standard rather than a vulnerability. No direct mitigation actions are required by defenders other than considering the adoption of AIUC-1 certified AI agents to reduce AI-related security risks. Enterprises should follow AIUC's guidance and leverage their audits and certifications to ensure AI agent security.

Summary

AIUC developed AIUC-1, an evaluation and risk standard for enterprise AI agents that tests AI systems against common security risks such as jailbreaks, hallucinations, prompt injections, anomalous behavior, and data leaks. The framework applies around 5,000 adversarial risk scenarios and performs quarterly audits to keep pace with evolving threat vectors. The standard is currently used to certify multiple AI agents including Cursor, ElevenLabs, Fin, Harvey, KPMG, Lovable, and UiPath. AIUC recently raised $40 million in funding to expand its audits, standards, and insurance offerings for frontier AI models, addressing the security and reliability bottleneck in enterprise AI adoption.

References

securityweek.com/aiuc-raises-40-million-to-certify-enterprise-ai-agents/ (opens in a new tab)

Community reviews

Loading reviews…

Description

AI analysisMachine-generated, updated 16 Sep 2026, 13:46 UTC

References

Community reviews