From insurability to bankability: why data centers must be engineered for systemic resilience

This surge in demand shows no signs of slowing. According to Goldman Sachs Research, the global power requirement for data centers is expected to increase by 50% by 2027 and soar up to 165% by the decade’s end.

This outpaced growth is putting capital at risk. Recent reports, such as from the Financial Times, indicate that some lenders are beginning to step back from data center financing due to insufficient insurance coverage.

At first glance, this appears to be an issue of limited insurance capacity. However, the reality is more nuanced. It’s about a lack of risk confidence – and it’s having an impact on how lenders think about financing.

Risk confidence is really a resilience challenge

Leading insurers like Zurich have been fast expanding their support for the data center sector. However, some lenders and investors still require or expect the market to provide full limit coverage or full replacement value. This reality is becoming a big challenge, as a single data center project can balloon into the tens of billions.

However, armed with a good understanding of the various risk scenarios and loss drivers, and supported by an Estimated Maximum Loss (EML) approach, this can help align and limit risk profile to loss potential. For developers, lenders, and investors to fully accept this and be confident with the approach, risk prevention and a resilience backbone need to be embedded at the core, with value articulation made clear.

Leading hyperscalers understand this reality. Increasingly, they accept only partial insurance, retaining the remaining risk on their own balance sheet or through captive arrangements. Many have built advanced internal resilience functions, giving them the confidence to manage any uninsured risk.

Why some infrastructure funds are stepping back

Infrastructure funds and private credit lenders often face different constraints. Few have in-house engineering teams capable of identifying vulnerabilities and addressing critical insurability issues. Without clear, independently validated evidence of resilience, partial insurance can feel like an unacceptable risk. As a result, capital sometimes walks away from even fundamentally strong projects – and this at a time when demand is red hot.

Based on Zurich’s industry analysis, the central issue is not a lack of insurability. Rather, it is that lenders and insurers need to see that resilience measures are in place to protect these new, complex operations and help ensure their high-stakes investments are risk-minimized.

Safeguarding an increasingly complex risk landscape

Today’s next-generation data centers are expected to deliver robust fire protection, redundancy, and operational controls – all to ensure seamless business continuity. However, the most significant vulnerabilities often sit beyond the walls of the center itself, in the broader ecosystems that keep them running. These systemic risks include:

• Grid reliability under single site loads of 1 GW or more
• Cooling water availability as climate conditions shift
• Exposure to natural catastrophes, including wildfire smoke, flooding, wind, and extreme heat
• Pathways to cascading failures (e.g., power loss → cooling disruption → thermal runaway)
• Non-damage business interruption, which can be more severe than physical losses

Many of these risks are systemic and interdependent, and by their nature only partially insurable. For this reason, the industry is moving from “full-value-cover” to extreme loss scenario models (such as probable maximum loss). Lenders are increasingly focused on the resilience of the broader ecosystem, not just the data center itself.

Operational continuity outages in focus.

When assessing this broader ecosystem resilience, catastrophic risk remains top of mind for any operator or investor. But interestingly, an analysis of major global data center outages over the past decade shows that the majority were not driven by an extreme, outlier type event.

While natural catastrophes pose a significant concern, most recent outages were primarily driven by electrical and power-related failures, which led to a number of cascading effects.

• 82% of events were driven by electrical or power system failures, including generator faults, protection misconfiguration, and power-quality disturbances
• Over 50% involved cascading failures, where an initial issue (e.g., power) propagated into cooling loss, system instability, or full-site outages
• More than one-third had clear human or procedural factors, including inadequate commissioning controls, unsafe interventions, or gaps in training

Across these events, outage duration ranged from hours to multiple days, with financial impacts reaching as high as several hundred million dollars per outage. What makes these challenges solvable is that many of the risks affecting data centers today are not ‘unknowns.’ They are recurring, measurable, and, in many cases, preventable. Operators we spoke with believe such events could have been avoided with stronger processes, design, and controls.

According to Doug Stohlman, VP of Construction Risk Engineering for Zurich Resilience Solutions, “We see a number of common threads with many of the major outage events that occurred, and each could have been minimized or altogether avoided with stronger design principles incorporated from the onset. For example, things like redundant and diversified cooling water paths, reduced energy density per room, and improved electrical isolation from grid disturbances, would have materially minimized losses in these cases.”

Resilience by design: the real line of defense

Insurance remains vital, but it should serve as the second line of defense. The financial impacts of a loss affect all parties, with the balance sheet bearing a significant share. That’s why “resilience by design” must become the first line of defense. Proactive risk engineering can integrate resilience principles from the start of any project, including:

• Location intelligence (climate, hydrology, and grid strength)
• Energy redundancy (dual feeds, on-site generation, and storage)
• Cooling strategies that are resilient to water stress
• Modelling of cascading failures across interconnected systems

When these elements are thoroughly documented, insurance capacity can be easier to deploy, and help lenders gain renewed confidence. A proactive, resilience by design approach means critical risks are addressed early in the build and design phase, help reducing hidden risks that could disrupt operations and stability. In turn, these multibillion-dollar investments are seen in a new, more favorable light.

Some of the approaches that risk specialists take into account include:

• Informed Site Selection and Design: Integrating resilience and risk mitigation in the planning phase via natural catastrophe modelling, power and water assessments. This can enhance both building and protection design.
• Making Risks Transparent and Quantifiable: Detailed risk analysis (fire, natural catastrophe, business interruption), together with financial indices and risk scores, can help leadership understand both probability and severity, while highlighting mitigation strategies.
• Reducing Estimated Losses: Ongoing assessments ensure the latest construction and protection measures are in place, and resilience indices tracking power, water, and critical infrastructure adequacy support continuity.
• Improving Insurability and Unlocking Bankability: Concrete recommendations on materials, emergency response, flood resilience, and more, can help turn uncertainty into clarity, making partial insurance a more manageable risk.

Insurability: the new gatekeeper for the AI infrastructure era

With data center expansion on a continued acceleration path (investments are projected to reach $6.7 trillion by 2030 according to McKinsey), insurance will remain a key discipline, rewarding projects built for resilience and filtering out those that are not.

Resilience will play an even greater role in underpinning insurability, which will in turn drive bankability – and this is ultimately the determining factor in what gets built. The winners will be those who see insurance as a strategic partner from day one, embedding “resilience by design” at the core of every development. As a result of a more strategic and holistic risk prevention approach, less capital will be exposed and put at risk, which will help continue the flow of investments into this market-altering segment.

Originally published in Commercial Risk on May 21, 2026