Monte Carlo sigma: a single AI campus can be a US$50bn replacement problem before capital is the constraint
Photo: Rsparks3, CC0, via Wikimedia Commons
Swiss Re Institute’s Sigma 3/2026 frames hyperscale campuses, geographic clustering (e.g. Texas/Virginia), shared grids, and multi-year high-voltage transformer lead times as converging accumulation drivers — engineering confidence, not spare capital, is the bottleneck. Energy and power underwriters writing generation, GSU, and campus power should map contingent BI and spare-transformer realism the same way property cat desks map wind zones. | Source: https://www.insurancebusinessmag.com/reinsurance/news/breaking-news/the-us50bn-datacentre-problem-reinsurers-must-solve-before-writing-the-risk-588893.aspx