AI Data Centre Insurance: The $200bn Infrastructure Opportunity
Swiss Re expects AI data centres and renewable energy infrastructure to generate around $200 billion in cumulative commercial insurance premiums by 2030, putting greater pressure on insurers and asset owners to understand how increasingly complex infrastructure behaves in the real world.
A new wave of infrastructure investment is creating an equally significant opportunity for the insurance industry.
Research published by Swiss Re Institute on 5 September estimates that AI data centres and renewable energy infrastructure could generate around $200 billion in cumulative commercial insurance premiums between 2026 and 2030.
Around $91 billion is expected to come from AI data centres, with approximately $111 billion generated by renewable energy projects. Swiss Re describes the wider investment cycle as a capital expenditure super-cycle, driven by the rapid construction of data centres, energy systems and other strategic infrastructure.
The scale of investment is already visible.
On 9 September, Google announced plans to invest $15 billion in AI infrastructure in Finland over the next two years, including three new data centres in the north of the country. The company has also entered into a 22-year agreement with Finnish energy company Fortum to purchase power from a nuclear plant.
For insurers, that creates an attractive new pool of business.
It also creates some unusually concentrated risks.
Billions of Dollars Can Sit Behind One Fence
AI infrastructure is becoming physically enormous.
Swiss Re says a large AI campus can cost billions of dollars to construct, with its value potentially doubling once high-value chips are installed. The resulting insurance requirements extend beyond physical property into business interruption, service interruption and liability.
The concentration matters because these facilities depend on far more than the servers visible inside them.
Power has to remain available. Cooling has to work. Network connectivity has to stay live. Physical access has to remain controlled. Fire protection, backup generation, batteries and monitoring systems all have to perform when called upon.
There are external risks too.
Swiss Re estimates that around 40% of US data centre capacity could sit in areas with significant tornado activity, with more than a quarter exposed to frequent large hail. Clusters of major facilities can therefore turn one regional weather event into losses across several valuable sites at once.
Renewable energy infrastructure creates a different set of dependencies. Wind and solar generation can rely on substations, transmission infrastructure, control systems, communications networks, specialist maintenance teams and supply chains for difficult-to-replace equipment.
Insurance increasingly has to account for the whole operating environment.
What Can An Underwriter Actually See?
Much of commercial insurance still depends on information supplied by the organisation seeking cover.
Questionnaires, engineering documents, certifications and control descriptions all contribute useful information.
Site-level evidence can reveal something different: how those controls operate in practice.
A backup generator may exist on the asset register. What matters during an outage is whether it starts, whether enough fuel is available, which systems it powers and how long the site can continue operating.
The same principle applies across physical and digital infrastructure.
Cyber policies may describe network segmentation. Engineers can establish how systems are actually connected.
Business continuity documentation may describe a fallback communications route. Testing shows whether people can use it when the primary connection disappears.
Security procedures may specify access controls. A site inspection can show how difficult those controls are to bypass.
For complex infrastructure, the gap between a declared control and an observed control can become financially significant.
Harry Geisler, Partner and Head of MEA at YAVA, says:
“Once you put billions of dollars of equipment into one site, the assumptions around that site really matter. An underwriter needs to understand what happens if the power drops, cooling fails, a network connection disappears or a physical control is bypassed. Those questions are much easier to answer when somebody has actually looked at the asset.”
Accumulation Risk Reaches Beyond Geography
Insurers have always considered whether multiple insured assets could be affected by the same event.
Technology infrastructure adds new forms of accumulation.
Several facilities can depend on the same electricity network. Multiple companies can rely on the same cloud region, network carrier or specialist equipment manufacturer. A cluster of data centres can share exposure to weather, water availability or a limited pool of engineering support.
Even individually resilient sites can therefore form part of a more fragile system.
This changes the questions worth asking during risk assessment.
How many critical systems rely on the same power source?
Where does connectivity leave the site?
Which equipment has a long replacement lead time?
Could one supplier failure interrupt several operations?
How much of the recovery plan depends on people or equipment that may also be needed elsewhere during a regional incident?
These dependencies can materially change the loss scenario.
What Should Insurers and Asset Owners Verify?
For high-value infrastructure, a useful site assessment should look across several layers at the same time.
Power and utilities: What happens when primary power disappears? How long can backup systems operate and when were they last tested under realistic load?
Technology: How are critical IT and operational systems connected? Where are the points through which disruption could move between environments?
Physical security: Who can access critical equipment, network rooms, control systems and power infrastructure?
Operational continuity: Which functions have to remain available during an incident and what are their dependencies?
Recovery: Who makes decisions when something fails? Which suppliers, engineers and replacement components are required?
Human controls: Do the people responsible for recovery understand the procedures and have they practised them?
The purpose is to build a clearer picture of how the asset behaves under pressure.
From Assessment to Remediation
Risk visibility becomes more valuable when it leads to action.
YAVA's Risk Intelligence work assesses external and physical risk, internal technology environments and human controls at the asset itself. Findings are then mapped against potential consequences and practical remediation priorities.
That might mean correcting a network weakness, improving physical controls, testing backup systems, redesigning communications or addressing infrastructure that has reached the point where failure becomes increasingly likely.
For insurers, stronger evidence can support underwriting and risk engineering.
For asset owners, it creates a prioritised view of where investment can reduce exposure.
For both, the conversation becomes grounded in the condition and behaviour of the asset.
A Growing Insurance Market Needs Better Evidence
Swiss Re's $200 billion estimate illustrates how quickly the commercial insurance opportunity around strategic infrastructure is expanding.
The underlying assets are becoming larger, more valuable and more interconnected at the same time.
That makes understanding dependencies increasingly important.
As investment flows into AI, energy and other critical infrastructure, insurers will need more than a view of what has been installed.
They will need confidence that the systems protecting those assets have been observed, challenged and tested.
And asset owners able to provide that evidence will enter the insurance conversation from a much stronger position.
FAQs
How large is the insurance opportunity from AI data centres and renewable energy?
Swiss Re Institute estimates that AI data centres and renewable energy infrastructure could generate around $200 billion in cumulative commercial insurance premiums between 2026 and 2030, including around $91 billion from AI data centres and $111 billion from renewable energy.
Why are AI data centres difficult to insure?
AI data centres concentrate high-value technology inside facilities that rely heavily on continuous electricity, cooling, connectivity, physical security and specialist equipment. Business interruption and accumulation risk can therefore become substantial.
What is engineering-led underwriting?
Engineering-led underwriting uses technical evidence from an asset and its operating environment to strengthen the insurer's understanding of risk. This can include site assessment, testing of controls, dependency mapping and analysis of potential failure scenarios.
What should insurers verify at critical infrastructure sites?
Useful areas include backup power, network resilience, physical security, equipment condition, recovery arrangements, communications, control-system dependencies and whether important fallback arrangements have been tested.
How does YAVA Risk Intelligence work?
YAVA combines on-site assessment with technology, engineering and operational analysis to identify weaknesses, understand their potential consequences and prioritise remediation.
Comments