Our Approach
The Problem
As climate change pressures mount, insurers, banks, and businesses face rising physical risks and billion-dollar losses. Traditional models calibrated on historical loss catalogs struggle as physical risk patterns shift — and the past becomes a weaker guide for the future.
Climate variables are changing in ways not captured by historical data. Frequency, intensity, and geographic distribution of hazards are all in flux. Models that rely on the past alone cannot adequately capture these emerging patterns.
Global climate models (GCMs) provide our best scientific estimate of how climate conditions may evolve. Yet their raw outputs are not designed for direct use in catastrophe models or physical risk assessment tools.
GCMs operate at coarse spatial resolutions and contain systematic biases that must be corrected before the outputs can be used in risk modelling. The scientific potential is immense — but bridging that gap requires specialist expertise.
Our Solution
Ötzi bridges that gap. We provide physical risk estimates for near- to mid-term horizons (+1 year) by correcting systematic biases, improving spatial resolution, and aligning outputs with the needs of decision-grade risk modelling.
Our AI simulator processes global climate model outputs and translates them into highly accurate, forward-looking physical risk estimates — at the spatial granularity, update frequency, and transparency that insurers, banks, and businesses require.
What Sets Us Apart
Physical risk estimates grounded in bias-corrected and downscaled climate model outputs — not historical averages. See the risks ahead, not just the risks behind.
Rapid processing enables frequent updates and rapid integration of new data sets and climate information — keeping risk estimates current as the climate signal evolves.
Risk estimates are grounded in the conducive conditions (physical variables) of hazards, making assumptions explicit and adjustments defensible — critical for regulatory scrutiny.
By modelling the climate system holistically, we capture event correlations and compound risks — critical for portfolios exposed to multiple concurrent perils.