South Korea plans 18.4 GW of AI data center capacity by 2035, driving more than KRW 1,000 trillion in investment. Insurers are still developing underwriting standards because fire and liquid cooling risks now drive most losses, leaving brokers to secure coverage before financing locks in.
Underwriting gaps and capacity constraints
Samsung Fire & Marine Insurance is preparing AI data center risk assessment guidelines with external experts, targeting a release in the second half of 2026. Hyundai Marine & Fire Insurance and Hanwha General Insurance are also reviewing their market approaches. The broader industry moved quickly after the Korea Fire Protection Association held a data center risk management seminar in July. Representatives from 11 non-life insurers attended to discuss assessment methods and claims handling.
International markets face similar pressure. Swiss Re estimates global insurance premiums tied to data centers will rise from $10.6 billion to $24.2 billion by 2030. Construction costs for a single site can exceed $20 billion, and insured values climb further once GPUs and other technology are installed. Swiss Re warned that traditional capacity may fall short of the limits lenders require for major projects.
How risk exposure is changing
AI facilities operate at substantially higher energy densities than traditional buildings and increasingly rely on liquid cooling and complex electrical infrastructure. Swiss Re cited industry data showing that fire accounted for 42.3 percent of traditional data center loss costs despite representing just 10.9 percent of loss events. Liquid-related incidents represented nearly a quarter of loss costs. Electrical supply failures drove 45 percent of outages, according to Uptime Institute data cited by Swiss Re.
These exposures cut across multiple policy lines. Large sites often insure buildings, equipment, and cooling systems separately, making it harder for carriers to track total exposure. Shared electrical, cooling, and fire-protection networks allow a single event to generate property and business interruption claims across several tenants and contracts.
Broker strategy and technical assessment
Limited operating history for next-generation AI sites forces carriers to lean on engineering assessments rather than historical claims. That creates pricing variation between markets. Brokers must engage specialist technical risk engineers at the initial submission stage. Financing terms should not dictate coverage structure.
Major carriers are already adjusting their programs. Aon raised its Data Center Lifecycle Insurance Program capacity to $5 billion in July. Marsh increased its Nimbus facility limits to $2.7 billion.
The Korea Insurance Research Institute flagged a widening gap between project values and available competitive capacity. Kim Jin-eok said, "Global insurance companies are rolling out dedicated programs that cover all stages from construction to operation under a single contract," and added, "As large-scale AI data center investment plans have been announced domestically, it is necessary to review the current fire insurance industry classification standards for data centers and the underwriting criteria for dedicated insurance." Professionals tracking AI for Insurance developments should build technical assessment capabilities into their standard workflows to manage shifting capacity and lender demands.
Why this matters for insurance professionals
Brokers and underwriters need to shift from reactive placement to proactive risk modeling. Early engineering input, clear aggregation mapping, and aligned policy sequencing will determine which projects get funded. Treat technical risk assessment as a prerequisite for submission, not a post-placement formality. Lenders will tie financing to coverage availability, so securing adequate limits requires upfront engineering validation and precise policy structuring.
Your membership also unlocks:
