The AI infrastructure gold rush is creating a problem that doesn’t get nearly enough attention: who insures all of it? AIG’s new CEO Eric Andersen just made clear that the explosion in data center construction is simultaneously the insurance industry’s greatest opportunity and its most complex headache.
Speaking at a recent S&P Global Ratings conference, Andersen described the data center buildout as the largest short-term opportunity for property and casualty insurers. But the risks involved, spanning construction, operations, cyber threats, liability, and even marine exposures, are pushing traditional insurers beyond what they can comfortably underwrite on their own.
A $24 billion problem by 2030
The numbers tell a striking story. According to Swiss Re, global data center insurance premiums sit at roughly $10.6 billion in 2026. By 2030, that figure is projected to exceed $24.2 billion. That’s more than a doubling in four years, driven by massive capital expenditures from hyperscale cloud providers racing to build out AI capacity.
Andersen, who took over as AIG’s President and CEO on June 1, 2026, succeeding Peter Zaffino, emphasized that the multifaceted nature of these risks means no single insurer can cover them adequately. The solution, he suggested, involves collaboration with alternative capital providers to bridge capacity gaps that traditional insurance markets can’t fill alone.
The haves and have-nots of insurance AI
Andersen also pointed to a growing divide within the insurance industry itself. Large insurers like AIG have the resources to invest in AI-driven tools that improve underwriting speed and accuracy. AIG’s own tool, called Underwriting by AIG Assist, is designed to help the company process risk assessments more efficiently, an increasingly critical capability when the risks being assessed are themselves AI-related.
Smaller firms, however, are struggling to keep pace. They lack the capital to build or acquire similar technology, and the gap is widening. In an industry where underwriting precision directly determines profitability, falling behind on AI adoption isn’t just a competitive disadvantage. It’s an existential risk.
Why this matters beyond insurance
Andersen’s comments also highlight the role of alternative capital in filling the gap. Catastrophe bonds, insurance-linked securities, and sidecar arrangements have long played a role in natural disaster coverage. Applying similar structures to data center risk could unlock new pools of capital, but it requires sophisticated risk modeling that the industry is still developing.
The projected doubling of data center insurance premiums to $24.2 billion by 2030 represents one of the fastest-growing segments in the entire insurance industry. For investors watching the insurance sector, Andersen’s framing is worth noting carefully. He called data centers a “great opportunity” while simultaneously acknowledging the industry is being pushed to its limits.