Over the past decade, data centre construction has expanded at an unprecedented scale, driven by explosive growth in cloud computing, digital services, and global connectivity. Hyperscale operators-such as major cloud and social media platforms-have led a wave of large-format developments, with each facility often spanning hundreds of thousands of square feet and housing tens of thousands of servers. This surge has transformed data centres from niche infrastructure into critical pillars of national digital economies, supporting everything from streaming and AI to e-commerce and enterprise platforms.

As a risk, data centres are mostly a very attractive proposition to insurers. They are site-constrained, built to high specifications, and (due to a number of developing factors) are now much more widely distributed geographically. The sector does not currently have an adverse construction loss history, and the figures involved, both in terms of capital expenditure and premiums generated are significant. Whilst project values are high, construction insurers allocate capacity on a Probable Maximum Loss (PML) basis, and the distribution of data centre hall structures generally reduces PML exposure. This makes the risks attractive to insurers and reduces the need to search for significant insurance capacity, even for some of the largest developments, unless there is material natural catastrophe exposure.

The only caveat to these plentiful capacity observations is the acute revenue-and therefore Delay in Start-Up (DSU)-exposure associated with many of these projects, given the significant commercial consequences of failing to commence operations on time. Whilst sufficient capacity exists for any one project, accumulation risk may increasingly be considered, and a major DSU loss remains the most likely factor to alter insurers’ currently positive outlook on the sector.

There are, of course, complexities in any project insurance programme design, and there are many delicate risk areas which we regularly see being overlooked by insurance advisors to data centre developers. We have seen serious errors in programme design that would result in developers paying significant premiums for DSU insurance without having the necessary triggers to ensure correct policy coverage. We have seen insurance programmes that do not reflect the reality of phased commercial operations dates, leaving substantial amounts of revenue uninsured. We have also seen basic, or even unchecked approaches to the review of risk allocation within project agreements with contractors, tenants, lenders, and utility providers, leaving developers heavily exposed to risk and potential financial indemnities. This is particularly salient at the point of handover, when SLAs (Service Level Agreements) with significant penalties for outages come into play. Without a diligent advisor exercising appropriate scrutiny at this critical stage, a significant loss is likely to result in a catastrophic commercial outcome for these projects.