AIUC raises $40 million Series A to expand AI agent certification to frontier models
Key Points
- AIUC raised a $40 million Series A led by Ribbit Capital, bringing total funding to $55 million, to extend its AI agent certification standard to frontier models.
- AIUC-1 tests agents against roughly 5,000 risk scenarios and is used by companies including Cursor, Harvey, ElevenLabs, and UiPath.
- The company plans to apply its audit, standards, and insurance model to frontier AI systems, echoing calls for independent third-party AI safety evaluation.
Artificial Intelligence Underwriting Company (AIUC) has raised a $40 million Series A led by Ribbit Capital, with participation from First Harmonic, the company announced on September 15, 2026. Combined with an earlier $15 million seed round led by NFDG, AIUC's total funding reaches $55 million. AIUC operates AIUC-1, a third-party testing and certification standard for AI agents, and the company said the new funding will extend that model from AI agents to frontier models.
What materially changed: extending certification from agents to frontier models
AIUC-1 currently tests AI agents against roughly 5,000 risk-and-attack combinations covering jailbreaks, hallucinations, and data leaks, tailored to each business type, according to the company. Certified agents are independently audited and recertified quarterly. AIUC said the new capital will be used to extend its audits, standards, and insurance model from the application layer, where AI agents operate, to the model layer, where frontier AI systems are built. AI tools assist with running and analyzing the tests, though humans verify the final results, co-founder Rune Kvist told The AI Insider.
Essential context: how AIUC's certification model works
AIUC-1 draws comparisons to existing industry certification models: the company's own materials cite Underwriters Laboratories, the organization that tests and certifies electrical products, while The AI Insider's reporting describes the approach as modeled on the cybersecurity standard SOC 2. More than 250 security and risk leaders from Fortune 1000 companies have shaped the standard through the AIUC Consortium, according to AIUC. The company was founded by Rune Kvist, previously Anthropic's first product hire, and Rajiv Dattani, previously a partner in McKinsey's insurance practice and chief operating officer at AI safety research organization METR. AIUC said frontier AI agent builders including Cursor, ElevenLabs, Harvey, KPMG, Lovable, UiPath, and Fin certify against its standard; The AI Insider's reporting separately named Cursor, Lovable, Harvey, and ElevenLabs among its customers.
Why it matters for enterprise AI buyers
AIUC's founders argue that risk assessment, not model capability, has become the main obstacle to enterprise AI adoption, pointing to agents that clear internal pilots but stall at security review. "Evidence of security and reliability is now the main bottleneck," said Rune Kvist, AIUC co-founder. The approach reflects a broader push for independent, third-party evaluation of AI systems: The AI Insider noted that AIUC's model echoes recent public calls from Anthropic CEO Dario Amodei for independent evaluators to oversee frontier AI safety. Ribbit Capital General Partner Nick Shalek said in the announcement that aligning AI builders, enterprises, security leaders, auditors, and insurers around a single standard was a core problem the company needed to solve.
What happens next
AIUC said it will use the new funding to extend its certification, audit, and insurance framework beyond AI agents to frontier models. The company has not disclosed a timeline for when model-layer certification will launch or which frontier model developers, if any, it is working with.