When AIG's Eric Andersen says AI data center coverage is 'maxing out' P&C providers, he is describing a structural failure, not a cyclical one. UK insurers regulated under PRA SS1/23 on third-party risk and FCA Consumer Duty PS22/9 are facing a genuine problem: they lack the data extraction and risk profiling infrastructure to underwrite novel hazards at scale. A data center's risk profile—power consumption patterns, thermal management, supply chain dependencies, cyber exposure—is fundamentally different from traditional commercial property. Yet most mid-market insurers are still relying on underwriters to manually review technical specifications and cross-reference coverage gaps. This is not a shortage of appetite or capital. It is a shortage of operational intelligence.

This story reveals where the insurance industry actually sits in its AI journey. Large carriers like AIG can absorb volatility and hire specialist teams. Smaller and mid-market insurers cannot. The result is market fragmentation: some capacity vanishes, premiums spike, and coverage remains patchy. Meanwhile, the legal, accountancy and financial services firms that depend on these insurers are exposed to hidden gaps. The narrative around 'AI capacity constraints' masks the real issue: insurers have not invested in the tooling required to extract, standardise and analyse the data that drives modern risk. They are trying to solve a data problem with capital and headcount.

Trovix's view is straightforward. Insurers do not need more AI chatbots or generative reasoning engines. They need document intelligence and data extraction at the point of underwriting. Trovix Sift extracts and structures unstructured technical data—specifications, compliance reports, audit trails—so that underwriting teams can model risk consistently and at speed. What Harvey and Legora do for legal research, Trovix Sift does for insurance underwriting: it turns unstructured narrative into structured insight. The alternative—waiting for humans to read and tag every data center specification sheet—is exactly what AIG is warning about. And that approach does not scale when risk profiles are novel and heterogeneous.

If you are a mid-market insurer, financial services firm or corporate buyer of data center coverage, act now. First: audit your underwriting data workflows. If your team is still manually extracting key risk factors from technical documents, you are burning capacity on a task that should be automated. Second: do not assume your broker or direct panel can fill coverage gaps for you. They cannot. You need visibility into what your insurer actually knows about your asset's risk profile before you sign on. Third: demand structured data from your insurers—not just quotations, but the underlying risk factors and exclusions in machine-readable format. This forces transparency and reduces disputes later. The firms that move fastest on this—law firms advising on AI infrastructure finance, accountants advising on insurable value, insurers building modern underwriting stacks—will have enormous competitive advantage over the next 18 months.

Source: Bloomberg News