AI agent certification startup AIUC raises $40M to begin auditing frontier models
SiliconANGLE Duncan Riley ● Covered by 71 sources
AIUC just raised $40M to audit frontier AI models, not just the agents built on them. The bet is that trust, not raw capability, is what’s blocking adoption now.
Based on reporting by SiliconANGLE, Duncan Riley — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
AI agent certification startup AIUC has raised $40 million to push its work up the stack, from the agents companies deploy to the frontier models those agents depend on. Until now, the company had focused on certifying software agents built on top of large models. Now it wants to inspect the models themselves.
The pitch is simple: enterprises aren’t waiting for AI to get smarter, they’re waiting for someone to sign off that it won’t embarrass them, leak data, or wander off script. Rune Kvist, AIUC’s co-founder and chief executive, said many companies already have agents that made it through pilots and then got stuck in security review. In his telling, proof of security and reliability has become the real bottleneck.
AIUC’s standard is called AIUC-1. Certification means running an agent through about 5,000 combinations of risk and attack, tailored to the kind of business using it. The tests look for jailbreaks, hallucinations and data leaks. Each agent is audited on its own and then recertified every quarter as attack methods evolve. Much of that testing is handled by AI agents, with humans checking the final audit.
The company has already found a market for that approach. ElevenLabs used certification in February to get what it called first-of-its-kind insurance for AI agents, including coverage for a voice agent that gives a customer bad information. Cursor maker Anysphere, legal AI company Harvey, KPMG, Lovable, UiPath and Fin are also on the list of certified users. AIUC says a consortium of more than 250 security and risk leaders from Fortune 1000 companies is helping shape the standard and push it inside their own shops.
Ribbit Capital led the Series A, with First Harmonic participating. Nat Friedman’s NFDG led a $15 million seed round when AIUC launched in July 2025. With this latest deal, investors have put $55 million into the company in total. The new money will help extend its audit and insurance work to frontier models, which AIUC now treats as the next place where trust gets stuck.
My take — AI-written commentary, not fact-checked reporting
This is what AI adoption looks like once the demo glow wears off: paperwork, insurance, and a lot of nervous people in security reviews. The real winners may be the firms selling trust, not the ones selling bigger models. That’s a very old software story in new clothes, and somehow still not glamorous enough for the people writing the cheques.
Read more about this at: SiliconANGLE