Underwriting Superintelligence: Backing Agents you can Sue — Rune Kvist, AIUC

AIUC, the Artificial Intelligence Underwriting Company, has raised a $40 million Series A led by Ribbit Capital and First Harmonic, cofounder Rune Kvist said in an episode of the Latent Space podcast with hosts swyx and Vibhu.

Kvist said AIUC is working with companies including Cursor, Harvey, Lovable and ElevenLabs. AIUC-1 is described as a standard for AI agent security, safety and reliability. In the episode, Kvist discussed how AI agents are stress-tested for jailbreaks, hallucinations and data leakage.

He traced his path into AI to reading the Scaling Laws paper late in 2021, after selling his first company, an edtech business. He said the paper convinced him that capital would understand predictable returns from larger models, and he moved to San Francisco to find the people who wrote it, joining Anthropic when it had around 40 people and later meeting Dario. He described early questions at the lab about whether to deploy models and how to engage with the world.

Kvist said the hypothesis behind AIUC's seed round was that risk would hold down adoption, which he called speculation at the time but now fact. The episode also covered the trust gap between governments and frontier labs, AI-enabled cyber and biological risks, AI liability and monitoring, copyright as a risk to insure, and whether AI engineers should be certified. It referenced the Air Canada chatbot case and a scenario in which a $20 Cursor subscription contributes to a $200 million plane crash. Kvist discussed AIUC's roadmap across agents, frontier models, robotics and red teaming, and said the labs can never fully serve as their own watchdogs.