The AI infrastructure buildout could generate about $200 billion in cumulative insurance premiums through 2030, according to a Swiss Re Institute report published Saturday. The forecast splits between $91 billion from AI data centers and $111 billion from the renewable energy projects feeding them, opening a major new line of business for commercial insurers at a time when the industry is searching for growth outside stagnant traditional lines.
Where the premium dollars will flow
Property insurance is expected to dominate the data center opportunity, accounting for $49 billion of the $91 billion total. Engineering coverage follows at $18 billion, with liability at $10 billion, credit and surety at $9 billion, and marine at $5 billion. Swiss Re said these figures represent a baseline - cumulative data center premiums could reach $105 billion under a high-capex scenario or fall to $80 billion if spending slows.
The market is already moving. Aon plc increased its Data Center Lifecycle Insurance Program to $5 billion in July, after raising capacity to $3.5 billion earlier this year. The expansion signals how insurers are scaling up to handle larger, more complex data center risks.
The scale of exposure is shifting
Some AI data center campuses now carry replacement values of up to $50 billion, Swiss Re said. That concentration of value at a single site changes the underwriting math. The five largest U.S. hyperscalers are expected to spend more than $800 billion on AI-related capital expenditure in 2026, while global data center capex is projected to exceed $1 trillion.
Nvidia has said it is working with financial institutions to help mobilize more than $500 billion in third-party capital for AI infrastructure. The capital flows are pulling insurance demand along with them, but the nature of coverage shifts over a facility's life. Engineering, marine cargo, and credit and surety are more important during planning and construction. Once operations begin, property, liability, and business interruption coverage take over.
Risk accumulation without a loss history
The report identifies four sources of risk accumulation: single-site exposures, geographic concentration, supply-chain dependencies, and shared physical and digital networks. Data centers often share suppliers, power systems, telecommunications, and cloud infrastructure. A single disruption could hit multiple insureds and lines of business simultaneously.
Insurers face a data problem. Many of these facilities have limited operating histories, making it difficult to estimate loss frequency or severity. Swiss Re said the issue is not a shortage of insurance capital but insurability. "Layered insurance programs, syndication, reinsurance and alternative capital can help spread those risks," the report noted.
Why this matters for insurance professionals
The $200 billion opportunity is not evenly distributed across the industry. Underwriters who can model concentrated tech risk, brokers who understand data center supply chains, and reinsurers with appetite for large single-site exposures will capture disproportionate share. The firms that build credible loss models for assets with thin claims histories will have a pricing advantage. For claims and risk engineering teams, the buildout means developing expertise in electrical systems, cooling infrastructure, and tech supply chains that most carriers do not yet have in house. This is not a distant trend - the capacity increases from Aon and the capex projections from hyperscalers show the market is forming now.
Professionals tracking how AI for Insurance reshapes underwriting and risk selection may find the data center buildout becomes a live case study in the next three years. The capital flowing into AI infrastructure is also reshaping AI for Finance as insurers and asset managers deploy funds into physical assets tied to computing demand.
Your membership also unlocks:
