Aon has added $1.5 billion of first-party capacity to its Data Center Lifecycle Insurance Program, pushing the ceiling to $5 billion and signaling that the brokerage expects the hyperscale build-out to keep demanding larger risk-transfer tickets. The expansion, announced Monday, arrives barely a year after the program launched in July 2025 with a $3.5 billion limit, a pace that suggests the insurance market is still scrambling to catch up with the capital intensity of AI infrastructure.
The program bundles construction all-risks, delay-in-startup, property damage and business interruption into a single tower backed by a panel of A-rated Lloyd's and company-market insurers. Cyber and technology errors-and-omissions coverage sits at $400 million, project cargo at $500 million, and third-party liability at $200 million outside the United States and $100 million within it. A separate terrorism facility adds up to $1 billion. Those sub-limits have not changed; only the first-party aggregate has grown.
Joe Peiser, chief executive of Aon's Risk Capital unit, framed the move as proof that the firm can help clients access capital, manage risk, and scale with confidence. The statement also introduced a suite of advisory services, climate risk, environmental solutions, security consulting, operational resilience, that Aon calls its Reliable by Design approach. In practice, that means the broker is now selling engineering opinions alongside the policy, a bundle that makes the coverage easier to underwrite and harder for clients to unbundle.
The counterpoint is that $5 billion still looks modest next to the multi-billion-dollar price tags of individual hyperscale campuses now in development. A single large language model training cluster can consume that much capital before the first GPU is racked. Insurers have historically been reluctant to concentrate limit on a single site, which is why the program relies on a panel rather than one balance sheet. The expansion shows appetite is widening, but it also highlights how much capacity remains fragmented.
What matters next is whether the advisory layer actually reduces loss frequency or simply creates a stickier distribution channel. If Aon's risk engineering can demonstrably lower construction delays and operational outages, the program becomes a genuine underwriting tool rather than a marketing wrapper. The next capacity increase, and the loss experience that justifies it, will answer that question.