Buying more time

Founded: 2019

HQ: Hamilton, Bermuda

AI for fire reinsurance.

Big Picture

Reinsurance, or insurance for insurance companies, is a 630-year-old industry that still uses literal rubber stamps and relies on the past to underwrite the future. Climate-fueled disasters that depart radically from the historical record are threatening the industry’s viability in many sectors. The result has been falling returns to reinsurance that risks leaving insurance companies without a safety net.

How it Works

Kettle uses deep learning to price reinsurance more accurately and granularly. Their algorithms divide a given area into a grid consisting of millions of smaller squares. It then calculates the probability of ignition for each of these squares, followed by a spread model that predicts which properties in an area are most likely to be affected. The result is a highly detailed picture of fire risk at an asset level.

Unfair Advantage

Through massive volumes of data and clever models, Kettle has built an algorithm that is 22% more accurate than industry standards. It draws on more than seven billion lines of data from satellites, weather, and real-time fires to run three tredecillion calculations per run. The result is that they can price risk in areas and for properties that a more superficial analysis would treat as too risky to cover.


Percent accuracy

in top 20% highest wildfire risk areas


Andrew was previously a VP of Digital Products at insurance underwriter Argo Group, where he built $72M+ in new product revenue.


Nathaniel was previously the CEO of Ushahidi and before that he was the first Chief Data Officer at USAID.

Follow: @natpmanning

This startup can predict where the next California fire will start

Fast Company

Fintech’s Wildfire Fighter: This Reinsurer Is Using A.I. To Make California’s Insurance Affordable