Mentioned:AIUCRibbit CapitalFirst HarmonicAnthropicMETR$40M · Series A
AIUC, founded by early Anthropic employee Rune Kvist and former METR COO Rajiv Dattani, raised $40 million in Series A funding to develop third-party audit and certification standards for AI agents. The company has built the AIUC-1 standard and a testing suite with 5,000 tests to validate agent safety for enterprise, government and military deployment.
What's happening
- AIUC announced a $40 million Series A led by Ribbit Capital with participation from First Harmonic, bringing total funding to $55 million including a prior $15 million seed.
- The company developed AIUC-1, a certification standard modeled on SOC 2 cybersecurity practices, and assembled a consortium of 250 security and risk leaders to define audit criteria.
- AIUC operates a testing suite of 5,000 tests that run agents through scenarios involving jailbreaks, hallucinations and data leaks, producing a 100-page safety report with human verification of final results.
- Named customers include Cursor, Lovable, Harvey and ElevenLabs; the company targets banks, hospitals, governments and militaries as buyers of certified agents.
Why it matters
- Defense and government agencies have adoption barriers tied to safety guarantees, not agent capability; AIUC certification removes a supply-chain gate blocking military and federal deployment of AI agents.
- Third-party audit and certification will accelerate agent procurement by institutions with compliance and accountability obligations, expanding the addressable market for frontier AI vendors.
- Independent safety validation mirrors the role that Anthropic CEO Dario Amodei recently called for, where embedded third-party evaluators verify safety at frontier labs, creating structural demand for AIUC's services.
Going deeper
- AIUC uses AI agents to run tests and analyze data, with humans verifying the final audit, creating a hybrid verification model that scales testing while maintaining human oversight.
- The consortium of 250 security and risk leaders directly shapes the AIUC-1 standard by specifying what they would ask when buying agents, ensuring the certification reflects buyer requirements.
- METR, Dattani's former employer where he remains a board member, focuses on agent performance testing for frontier labs; AIUC targets the enterprise and military buyer side, creating a complementary but separate market.
Financial impact
- The $40 million Series A raises AIUC's cash position by $40 million in Q3 2026, funding product expansion, consortium operations and go-to-market for defense and government segments.
- Revenue model is subscription or per-audit, not disclosed; timing and gross margin characteristics depend on whether AIUC recognizes revenue per-test, per-certification or on a recurring license basis, not stated in available text.
- As a private company, the Series A dilutes prior seed investors, including Nat Friedman's NFDG, Emergence, Terrain and Anthropic co-founder Ben Mann; post-money valuation not disclosed.
The intrigue
- AIUC's reliance on AI to execute and analyze 5,000 tests raises a second-order question: whether the testing methodology itself requires the same certification it applies to customer agents, and whether regulators will accept AI-driven audits of AI safety.
- Dattani's dual role as AIUC founder and METR board member creates potential alignment on safety standards but also raises questions about information flow and whether METR's frontier-lab testing and AIUC's enterprise audits will converge into a single industry standard.
The fine print
- No customer adoption contracts, contract values or revenue milestones disclosed; named customers (Cursor, Lovable, Harvey, ElevenLabs) may be early-stage or non-binding relationships.
- AIUC-1 standard and testing suite are proprietary; no statement on whether the standard will be submitted to NIST, DoD or other regulators for formal recognition as a defense procurement requirement.
Source: TechCrunch · Sep 15, 2026 · Drafted with Claude, reviewed by Industrial Base Alpha.