Analysis
Harbor emerged from stealth with $10 million in seed funding led by Base10, Abstract and 8VC, with Restive, NFX, Innovation Endeavors, Kibo Ventures, Masia and MVP Ventures also participating, the company said this week, according to Pulse 2.0. Harbor builds insurance and risk infrastructure specifically for companies deploying AI agents inside their operations.
The product identifies which AI agents are actually running inside a company, then assesses their operational, security and regulatory risk. It can block a decision before it executes if that decision would violate a company's own rules. Harbor then uses that same risk-measurement layer to underwrite and insure whatever exposure remains after those controls are applied -- combining monitoring, controls and insurance into a single model rather than three separate vendors, as is typical today.
Betting On A Narrower, Deeper Wedge Than Existing AI Insurers
Pulse has previously covered Corgi, the AI-native insurance carrier that underwrites broad commercial coverage -- D&O, cyber, general liability and AI liability -- for tech startups generally, and which hit a $4 billion valuation earlier this year across three rounds. Harbor is making a narrower bet: rather than insuring a company's general AI exposure, it is trying to own the specific, technical layer of measuring and controlling what autonomous agents actually do, then selling the insurance as the output of that measurement. CEO Jesus Santos described the goal directly: "We're making these risks measurable, manageable and, most importantly, insurable. We're building a new category of insurance."
Harbor was founded by four people who met at Stanford, and the broader team includes former founders, International Mathematical Olympiad and International Olympiad in Informatics medalists, competitive programmers and former professional athletes, working across San Francisco, the rest of the US and Europe. Its first customers are banks and insurers across the United States, Europe and Latin America -- regulated industries where an AI agent's mistake can carry direct financial and compliance consequences.
The structural risk is that Harbor is both the referee and the insurer: it measures AI-agent risk with its own models, then carries the financial exposure on the resulting insurance product itself. That only holds up if Harbor's risk models are accurate; the company has not named a reinsurance partner to absorb losses if its underwriting turns out to be wrong, and it has not disclosed prior funding, headquarters location or founding year.
Harbor has not named a reinsurer to share that tail risk, and its first banking and insurance customers across three regulatory regimes haven't completed a single underwriting renewal yet.