A single AI data-centre campus can cost up to US$50 billion to replace. That figure, from Swiss Re Institute's Sigma 3/2026, is not a projection. It is a present-day underwriting reality the reinsurance market has barely begun to price.
The sigma report, released at the Rendez-Vous de Septembre in Monte Carlo, predicts global energy investment to reach US$3.4 trillion in 2026, with around US$2.2 trillion directed at renewables, nuclear, grids, storage, and electrification. The five largest US hyperscalers are projected to spend nearly US$800 billion on AI-related capital expenditure this year alone.
AI data centres and renewable energy infrastructure could generate around US$200 billion in cumulative premiums between 2026 and 2030, Swiss Re Institute estimates. The size of that opportunity depends entirely on the market's ability to manage what comes with it.
The sigma report identifies four structural drivers of risk accumulation now converging: large individual assets, geographic clustering, supply-chain dependencies, and shared physical and digital networks. Each reinforces the others, so a single disruption can affect multiple policyholders and lines of business simultaneously.
The geographic concentration is already measurable. Texas and Virginia alone account for more than 40% of current and planned US data-centre capacity. More than a quarter of that capacity sits in hail-exposed zones, and around 40% falls within areas subject to at least three tornado days annually.
Taiwan presents a different version of the same problem. Around 88% of semiconductor fabrication plants there are located in extreme to very extreme seismic-risk zones. A major seismic event would carry downstream consequences across multiple industries and geographies, given Taiwan's position in global supply chains.
High-voltage transformers can carry lead times of multiple years. A damage event that delays replacement can, therefore, extend business interruption losses well beyond the physical damage period. Standard policy structures are not designed to capture those contingent exposures.
"We are seeing the digital economy become a real economy," said Gianfranco Lot, Swiss Re's chief underwriting officer for P&C Re. "AI needs data centres, power grids, and increasingly complex infrastructure - and all of it needs insurance. That creates growth opportunities across multiple lines of business, but also significant risk concentrations. The deployment of capacity will depend on our ability to understand and manage those, and getting paid for the associated tail risk."
Swiss Re Institute is direct on where the bottleneck lies. The principal constraint is not the availability of insurance capital. It is the ability to deploy that capital confidently against exposures with short operating histories, extreme loss potential, and complex interdependencies.
The numbers make the gap concrete. S&P Global Ratings estimates total insurable values for a single hyperscale campus at US$20 billion to US$30 billion. Per-risk limits from large carriers sit in the low single-digit billions, so full-value coverage requires stacking multiple carriers in layered structures.
Data centres compound the problem further. The operational phase introduces complex coverage demands, with property, business interruption, contingent business interruption and liability exposures layered across the same asset. In some cases, financial losses from a service interruption can exceed the physical damage itself.
Large facilities are frequently presented to insurers through separate programmes covering buildings, equipment and power plants independently. That fragmentation makes overall accumulation difficult to track. Gallagher Re's recently established digital risk practice was created partly to address that gap, pulling AI liability, data-centre risk engineering and cyber expertise into a single advisory function.
Jérôme Haegeli, group chief economist and head of Swiss Re Institute, set out the broader stakes. "A new investment era is taking shape, with unprecedented amounts of capital flowing into the infrastructure that will power future economic growth," he said. "This also concentrates more value and creates new dependencies across power systems, supply chains and digital networks. Insurance is essential to making these investments resilient and financeable."
Swiss Re Institute argues that three capabilities are required before reinsurers can deploy capacity at scale: engineering-led underwriting, improved accumulation modelling, and disciplined accumulation management. Together, these raise confidence in the underlying risk, which is the prerequisite for broader market participation.
Risk distribution across insurers, reinsurers, and capital markets spreads large exposures across multiple balance sheets, keeping assets insurable at a price that supports continued investment. Without it, the largest projects face either uninsured gaps or prohibitive pricing that undermines the build's economics.
Insurance markets have adapted to emerging risk classes before, with nuclear power and cyber as the sigma report's own examples. As the reinsurance market works through what deploying capital against new accumulation challenges actually requires, the capex super-cycle is the next test of that adaptability. The first renewals that price these risks credibly will show whether the market is ready.
