Team
80Climate scientist who stayed for the sales calls, plus a CTO from mapping infrastructure. Rare combination, still a small bench.
This part of the screen is weighted at 30 percent.
A geospatial model that prices physical climate risk for insurers and grid operators.
Climate scientist who stayed for the sales calls, plus a CTO from mapping infrastructure. Rare combination, still a small bench.
This part of the screen is weighted at 30 percent.
Insurers already buy catastrophe models. The budget is not hypothetical. Switching costs are political.
This part of the screen is weighted at 25 percent.
€1.2m ARR, three paid pilots converting slowly. The pipeline is wider than the revenue, which is either healthy or a warning.
This part of the screen is weighted at 25 percent.
Labelled loss data from two carriers and a grid operator. That corpus is the asset. It is not yet exclusive.
This part of the screen is weighted at 20 percent.
Screened 7 Oct 2026Pre-written. The replay is simulated.
Cartograph sells a risk surface, not a dashboard. A carrier or a grid operator asks what a flood or a heat week does to a book of assets. The model answers with a number their existing vendor also answers, and the diligence is whether Cartograph is righter often enough to switch.
Two European carriers are rebidding catastrophe tooling in the next two quarters. Cartograph is in one process and adjacent to the other. Kiln’s buyer made the introduction, which is a real reference, not a logo slide.
They fine-tune on loss events the customers provide, on top of public satellite and weather series. Outputs land in the tools actuaries already use. The interface is deliberately dull. The value is the delta versus the incumbent model on the last five years of claims.
Annual licence, €1.2m ARR, gross margin 82% because compute is bursty and someone else owns the satellites. Sales cycles are two quarters. A €5m cheque buys us a seat while those rebids resolve. It does not buy a winner.
Stay in diligence through the first rebid. If Cartograph is a finalist and the backtest beats the incumbent on the customer’s own claims, bring it to IC. If they are a ‘nice second model’, pass without drama.