AIUC raises $40 million to certify enterprise AI agents

AIUC has raised $40 million to build out a certification scheme for enterprise AI agents, SecurityWeek reported on 16 September 2026. The company offers a standard for AI systems and tests them against risks including jailbreaks, prompt injections and unauthorised actions, according to the report.

Those three categories map closely to the problems security teams keep running into when they put agents into production. A jailbreak is an input crafted to make a model ignore the rules its operator set. Prompt injection is the related trick of hiding instructions in data the agent reads, such as a web page, a document or an email, so the agent follows the attacker's text instead of the user's. Unauthorised actions are the consequence: an agent with access to a mailbox, a payment system or a code repository doing something nobody approved.

The pitch for certification is that buyers currently have no common yardstick. An enterprise evaluating an agent vendor has to design its own tests, or take the supplier's word for it. A shared standard, with an outside party running the tests, is meant to give procurement and risk functions something they can compare across suppliers. SecurityWeek describes AIUC's product as exactly that: a standard, plus testing against a defined risk list.

How much assurance such a scheme can offer remains an open question, and the source material does not settle it. Certification is a point-in-time exercise, while models, system prompts and tool permissions change constantly. A pass against a fixed list of jailbreak and injection techniques does not prove an agent will resist a technique invented next month. That is a familiar limitation from other parts of security, where certificates and audit reports coexist with breaches.

What the funding round does indicate is where money is going. Investors are betting that enterprises will pay for third-party evidence that their agents behave, rather than building that evidence in-house. The size of the round, $40 million, suggests the backers expect demand from large buyers rather than a niche compliance product.

SecurityWeek did not report further detail on the scheme's scope, the identity of AIUC's investors, or which customers have been certified so far.