Willis has urged data centre owners, developers, builders, operators and investors to rethink how they buy insurance, warning that many organisations may be buying more cover than their actual exposure requires.
The warning comes as investment in digital infrastructure rises and AI-related demand drives the construction of larger facilities with more concentrated risks.
Willis said the global insurance market can provide up to USD $15 billion in cover for large-scale data centre risks when needed. However, it argues that the key issue is not the amount of insurance available, but whether buyers have a clear enough view of their real exposure across design, construction and operations.
Insurance buying in the sector has often focused on building larger insurance towers as projects grow bigger and more complex. Willis argues that a risk-led approach, based on detailed analysis, can help companies avoid unnecessary spending while shaping decisions on resilience measures and capital allocation.
Digital infrastructure projects face a wide range of risks that go beyond the value of the physical asset itself. Exposures can differ sharply depending on site location, access to power, construction methods, operational resilience, supply-chain reliance, climate threats and cyber weaknesses.
That matters because larger, more interconnected facilities can create greater concentrations of risk, including business interruption and dependency exposures. A single outage or physical disruption can affect more than one building if critical services such as energy, water and cooling are disrupted.
Lay See Ong, Digital Infrastructure Industry Leader, Asia, at Willis, pointed to that shift in the regional market.
"The rapid growth of AI-enabled digital infrastructure is prompting organisations to reassess key assumptions regarding asset concentrations, business interruption exposures and dependency risks, especially in Asia. As facilities become larger, more power-intensive and increasingly interconnected, risk quantification is becoming an important tool in determining appropriate resilience investments and insurance requirements," said Ong.
Risk analysis
A more detailed understanding of exposures can support changes not only to insurance structures, but also to the underlying design of facilities. Willis referred to an eight-point digital infrastructure risk framework it uses to assess these issues, although it did not set out the full model in detail.
Its position is that spending on risk engineering and resilience can, in some cases, reduce overall exposure more effectively than simply raising policy limits. In practice, developers and owners may benefit from investing in protective measures early in a project rather than relying on additional insurance after construction.
Examples include flood protection, stronger wind resistance, seismic design measures, adaptations for heat and drought, wildfire mitigation and blast resistance. These can be built into projects early if natural hazard and climate risks are assessed during development.
Willis also argued that cost-benefit analysis can help organisations decide which resilience investments are worthwhile. Such assessments can also support discussions with insurers, lenders and investors when companies want to show that their risk profile is better understood and managed.
Financing effects
Willis said this type of analysis has affected financing as well as insurance outcomes for some clients. It said the approach had helped improve credit and financing terms, including stronger S&P ratings and better terms for future development projects. It also supported lower insurance limits where the underlying risk proved lower than first assumed.
The same work also helped clients improve resilience to natural hazards and strengthen continuity planning by estimating likely downtime from major catastrophe events and identifying specific mitigation steps.
Alastair Swift, Head of Global Specialties and the Global Digital Infrastructure Group at Willis, said insurance limits alone do not determine protection.
"Buying more insurance is not always the same as being better protected. When risks are properly modelled, understood and mitigated, clients can build more efficient, resilient insurance programs that reflect their actual exposures. This is especially important where lenders and equity partners expect robust protection; a more tailored approach can often deliver greater value," said Swift.
Market shift
Willis said clients should move away from capacity-led buying and towards decisions based on data and quantified exposure. It outlined several steps, including modelling realistic loss scenarios, measuring risk across the full life cycle of an asset, building resilience into projects from the outset, and assessing dependencies tied to power, water, cooling and continuity planning.
Robust, verifiable data can also strengthen discussions with insurers and financial stakeholders by helping them judge whether cover matches the actual scale of risk. That is becoming more relevant as lenders and equity investors scrutinise whether insurance programmes are sufficient without being excessive.
In Asia, Willis said it is seeing more detailed discussions among owners, investors and lenders about whether programmes are appropriately sized. Rather than relying on market convention, these parties are increasingly asking for evidence that insurance levels and resilience measures reflect a project's specific risk profile.
"Across Asia, we are seeing increasingly sophisticated discussions around insurance purchasing. Owners and investors are asking whether they are buying more insurance than necessary, while lenders are seeking greater confidence that programmes provide adequate protection. The common thread is a desire to make decisions based on data rather than convention.
"Client insights show that sharper risk quantification can help stakeholders across the digital infrastructure ecosystem make smarter decisions, avoid unnecessary insurance spend and ensure insurance and risk programmes remain appropriate as digital infrastructure becomes larger, denser and more interconnected," added Ong.