Four points on risk, insurability, and AI from a recent interview with Andreas Berger, CEO of reinsurer Swiss Re. who makes some solid points. My take, however,: AI and climate change are sharply increasing cross-line correlations, and the industry thinks it can price and "big data" its way out. It likely can't.
1. Catastrophe losses are mostly an asset-density story
Reinsurers have a more prosaic explanation for rising disaster losses than the public narrative. The main driver is exposure.
More people, houses, and infrastructure are being built in the highest-risk locations: coasts, wildfire corridors, and floodplains. More people are building blithely with AI and agentic systems. When events occur, the damage pool is simply larger.
Swiss Re notes insured catastrophe losses have exceeded $100B annually for six consecutive years, with $318B in total economic losses in 2024. Climate likely contributes to severity, but the structural trend is that capital keeps moving into exposed areas.
2. “Uninsurable” actually means “politically unaffordable”