AIUC raises $40M Series A to rein in rogue AI agents

TechCrunch reported on September 15 that AIUC, the Artificial Intelligence Underwriting Company, has raised a $40 million Series A led by Ribbit Capital, with First Harmonic participating.

The article frames the startup's goal plainly: finding a way to rein in rogue AI agents. Its founding team pairs an early Anthropic hire with a former METR chief operating officer, a background rooted in AI safety and capability evaluation.

The word underwriting in the company's name suggests the approach: rather than patching a misbehaving agent after the fact, price the risk and draw the lines of liability before deployment starts.

That direction has become pressing because agents are changing shape. When a model stops answering questions and starts executing tasks, its mistakes are no longer only bad text; they become transactions, tickets and changed system states with measurable economic consequences.

The round is modest by current standards, but the involvement of Ribbit Capital as lead, with First Harmonic participating, signals that third-party management of agent risk is being treated as a market of its own rather than a feature of another model vendor.

What remains unproven is the delivery model. Insurance, guarantees, audits or runtime supervision: none has yet been accepted by both enterprise buyers and regulators, and this company is betting that pricing agent risk is the piece that gets adopted first.

The next signals to watch are the first named customers and the concrete product form. Who is willing to pay for agent risk cover, and whether such arrangements hold up when a real deployment goes wrong, will decide whether underwriting becomes infrastructure for agents or a niche.