Startup founded by Anthropic and METR alumni raises $40M to audit AI agents

AIUC, a startup that audits artificial intelligence agents for enterprise buyers, has raised a $40 million Series A led by Ribbit Capital, with participation from First Harmonic. The round brings the company’s total funding to $55 million, following a $15 million seed round backed by Nat Friedman’s fund NFDG, along with Emergence, Terrain and Anthropic co-founder Ben Mann. The company, whose full name is Artificial Intelligence Underwriting Company, was founded by Rune Kvist, an early employee at Anthropic, and Rajiv Dattani, the former chief operating officer at AI safety research group METR. The two are brothers-in-law; Kvist is married to Dattani’s sister.

Kvist said AI models become harder to control as they get smarter, not easier, which is why enterprises hesitate to deploy them. Banks, hospitals, governments and militaries do not skip AI deployment because a model lacks capability, he said. They skip it because they have made commitments to their own customers about what a system will and won’t do, and no one can currently guarantee that behavior.

A SOC 2 for AI agents

AIUC built its standard, called AIUC-1, on the model of SOC 2, a widely used framework that audits how software companies protect customer data. To write the standard, AIUC assembled a consortium of about 250 security and risk leaders, the people who buy AI agents inside enterprises, and asked them what they look for and what questions they want answered before a purchase. That feedback becomes the basis for testing. AIUC runs an agent through roughly 5,000 tests covering jailbreaks, hallucinations and data leaks, then produces a report of about 100 pages showing where the agent performs safely and where it does not. AI agents run the tests and AI tools analyze the resulting data, but humans verify the final audit, according to Kvist. AIUC lists Cursor, Lovable, Harvey and ElevenLabs among its customers.

Dattani was METR’s chief operating officer from 2024 to 2025 and remains on its board. METR performs similar evaluations for frontier AI labs, though its work has focused mainly on whether agents can complete tasks reliably rather than on safety behavior. METR was among the independent organizations OpenAI brought in to investigate its Hugging Face security incident. The launch follows recent comments from Anthropic chief executive Dario Amodei, who has called on the AI industry to slow the pace of frontier development, citing a rise in reports of models behaving badly. Amodei floated requiring frontier labs to use embedded third-party evaluators to observe and verify safety, naming METR as one candidate. AIUC is not proposing to place itself inside customer companies. Instead, Dattani said, the company hands enterprises an independent assessment: where an agent can be trusted, where it can’t, and what to know before buying it.

Watch for AIUC to publish results from AIUC-1 audits publicly, and for whether other AI-safety evaluators adopt a similar third-party certification model as enterprises face more pressure to prove their AI agents behave as promised.