More than nine in 10 leaders across the data center value chain experienced a material disruption in the past five years, yet most are not prepared for what comes next. A new global study found that fewer than a third run compound-shock simulations to test how those disruptions might travel through interconnected systems.
The research was published by Economist Enterprise and is sponsored by global property insurer FM. It draws on a survey of 1,800 senior executives across 21 countries. The report found that risk accumulates at the level of the interconnected system, but preparedness remains concentrated at the level of the individual firm.
"As organizations scale, the challenge is not simply managing individual risks, but understanding how cyber, energy, climate and supply-chain disruptions can interact across increasingly interconnected systems," said Christopher Dempsey, senior vice president at FM Intellium. "Resilience needs to keep pace with growth and inform critical decisions early."
Investment in resilience is widespread but uneven. Three-quarters of respondents said their organizations are funding resilience initiatives, and around two-thirds reported board-level oversight. When forced to make trade-offs, more than 80% chose growth, efficiency, and AI deployment over long-term resilience.
Seven in 10 respondents identified short-term financial targets as a barrier to strengthening preparedness. Cybersecurity is the most widely adopted measure at 91%, followed by physical protection of control systems at 88%, and energy demand management at 87%. Those figures show strong coverage of visible immediate threats.
Compound shocks receive far less attention. Only 12% of respondents use geographic diversification as a resilience strategy. Just 22% apply climate-informed criteria when selecting new facility sites.
The report identifies two structural features that make cascading failures hard to contain. The first is tight coupling: facilities, grids, and networks are linked closely enough that stress in one spreads quickly to the others. The second is concentration, where a small number of power markets, cable routes, and equipment suppliers hold a disproportionate share of global capacity.
For investors, the stakes show up directly in the data. Among those surveyed, withdrawal of investment or insurance coverage ranked as the second most important systemic risk after cyberattacks. That finding connects to unresolved questions about business interruption coverage at operational data center scale, where calculating outage loss in owner-operated facilities remains contested.
Most organizations are not structured for cross-value-chain coordination. Just 10% maintain structured crisis plans with all critical external partners. Almost two-thirds engage with only a select few, and a quarter rely on internal planning alone.
Only a third possess cross-functional crisis teams, and compound-shock simulations are conducted by fewer than one in three respondents. Without those exercises, organizations have limited visibility into how a disruption might travel through their broader network.
The workforce dimension adds a further layer of exposure. About 60% of data center operators globally said they struggle to recruit qualified staff, with direct consequences for incident resolution times. Around 35% of operators, the highest share of any stakeholder group, said they expect skills shortages to lead to infrastructure failures.
One in six investors identified workforce capability as a systemic risk, a narrower view than the operators actually running the facilities. The staffing constraint spans data center construction and operations alike, where demand for skilled trades continues to outpace supply.
Slower-moving risks are receiving the least attention despite their long-term impact on investment outcomes. Fewer than 5% of respondents ranked community opposition among their top five systemic risks. Barely a tenth cited technological obsolescence.
A Gallup poll published in May found that seven in 10 Americans oppose data centers in their area, and that opposition is moving beyond public sentiment. According to Data Centre Watch, which tracks local resistance to data center construction in the US, it contributed to delays or cancellations on more than 75 projects worth approximately US$130 billion in early 2026 alone.
The research modeled a scenario in which a rapid shift to ultra-high-density AI chips leaves legacy facilities unable to support next-generation workloads. Tenants migrate to purpose-built sites, asset values fall, and operators are left with stranded infrastructure. That long-cycle risk does not appear in standard operational stress tests but can determine coverage viability and investment returns.
Policy language in data center cyber accounts faces a similar lag, where coverage structures have yet to catch up with how risk profiles have shifted.
"The challenge is that each stakeholder in the value chain is more focused on securing their own facility or connection, whereas genuine resilience requires understanding how risks interact across the system," said Jonathan Birdwell, global head of thought leadership at Economist Enterprise. "Closing the gap between where risk lives and where it is managed will be a defining challenge for the industry over the next decade."
