Swiss Re: global premiums for insuring data centers will likely reach $20B-$30B/year by 2030; ~40% of US data-center capacity is located in tornado-prone areas
AI hyperscale data centers are exposed to a dizzying array of potential threats during construction and operation
Context & Ripple Effects
Data-center expansion was already becoming a capital-intensive buildout: global construction spending was tracking toward $250 billion annually in 2024, while forecasts called for 8,400 global facilities by 2030. The accompanying power requirement is also material, with data centers projected to account for nearly half of US electricity-demand growth through 2030.
Swiss Re adds physical-risk underwriting to that investment case. Its estimate makes insurance exposure a measurable constraint alongside the electricity and financing pressures surrounding AI infrastructure.
First-order effects
- Data-center developers and operators, particularly those with US capacity in tornado-prone areas, face a larger insurance cost and coverage-planning requirement as facilities are built and operated.
- Swiss Re and other insurers gain a defined growth market in underwriting construction and operating risks for AI data centers.
Second-order effects
- Lenders and investors financing data-center projects will have to incorporate insurability and premium costs into project economics, alongside the debt-heavy funding backdrop documented in 2025 data-center deal and debt activity.
- Site selection and facility design become more consequential for operators where exposure to tornado risk raises the cost or complexity of obtaining coverage.
Third-order effects
- If premiums develop toward Swiss Re's range, insurance capacity and physical-risk exposure will become part of what makes an AI data center financeable, not merely an operating expense.
- The AI buildout is moving toward a broader definition of compute execution risk, in which power availability, construction exposure, and catastrophe coverage jointly shape deployable capacity.
The trend: AI infrastructure is being financialized as a long-duration physical asset whose power, construction, and catastrophe risks must be priced before capacity can be financed.