AIUC raises $40M Series A to underwrite AI agents and frontier models

Artificial Intelligence Underwriting Company (AIUC) said it raised a $40 million Series A led by Ribbit Capital with participation from First Harmonic. Combined with a prior $15 million seed from NFDG, the San Francisco company says total funding is now $55 million to expand audits, standards, and insurance from AI agents toward frontier models.

Same-day company primary that venture money is going into audits, standards, and insurance so enterprises can deploy AI agents — and later frontier models — with a named trustmark, not another chatbot wrapper. A $40 million Ribbit-led Series A is a concrete FINANCE signal that “can we trust this agent” is becoming its own infrastructure layer.

Artificial Intelligence Underwriting Company, or AIUC, published “Series A to build the confidence infrastructure for frontier AI” on its company updates page on 15 Sep 2026. The hed on that page is “AIUC raises $40 million Series A from Ribbit & First Harmonic.” Authors listed: Rajiv Dattani and Rune Kvist. Dated Sep 15, 2026. Series A is early growth funding after seed money — not a first check, and not a later growth round. That company page is the filing event. These are company claims.

The round, as AIUC tells it: $40 million Series A led by Ribbit Capital, with participation from First Harmonic. None was printed.

Prior raise and total, still the company’s: a $15 million seed from NFDG. Combined with the Series A, AIUC says total funding is now $55 million. Seed is early venture funding — a first institutional check.

Mission, as the same page has it: build the audits, standards, and insurance that let enterprises deploy AI with confidence. The company slogan on the page is “Underwrite Superintelligence.” An audit here is an independent check against a written standard. Insurance here is coverage meant to sit on top of that standard, not a policy this desk reviewed.

Product, still company: AIUC-1 is the standard for AI agents. It tests how well agents hold up against risks such as jailbreaks, hallucinations, and data leaks, using 5,000 risk-and-attack combinations the company says are unique to each type of business. A jailbreak is a trick that gets an AI to ignore its safety rules. A hallucination is a confident-sounding answer that is made up or wrong. The company says it built the standard with 250-plus security and risk leaders. File AIUC-1, the 5,000 combinations, and the 250-plus builders as AIUC’s.

Company-listed names that certify against AIUC-1: Cursor, ElevenLabs, Harvey, KPMG, Lovable, UiPath, and Fin. The page says companies building widely used agents — including those names — certify against the standard, and that AI which writes production code, drafts legal work, and answers customer calls now holds the AIUC-1 trustmark.

Use of proceeds and hiring, company: the capital will help extend audits, standards, and insurance from agents to frontier models. Frontier models are the largest, most capable general AI systems labs are still racing to build. The company says it is hiring across all roles in San Francisco.

Company history lesson, attributed: AIUC compares the job to electricity, cars, and nuclear power — standards, audits, and insurance creating a trust flywheel. It names Underwriters Laboratories, or UL, as the example of insurers funding a lab to certify products after electricity burned down houses. Longer color on that flywheel stays in Sources.

Plain English for the rest of the card: Series A = early growth funding after seed. Seed = early venture funding. Audit = independent check against a written standard. Jailbreak = a trick that gets an AI to ignore its safety rules. Hallucination = a confident-sounding answer that is made up or wrong. Frontier models = the largest, most capable general AI systems still being built. Trustmark = a certification badge that a product passed a named standard.