Skip to content

Five Ways the AI Race Could Impact Cat Bonds

You can’t read the news today without hearing about AI and its far-reaching impact on our daily lives. From the way we work to how we interact with information and even how AI is reshaping entire industries, it seems like nothing is immune to AI’s influence. Cat bonds are no exception.

As tech companies race to build data centers to support growing demand, it’s important to take a look at how this construction boom could potentially impact cat bonds. Here are five things we’re keeping an eye on.

  1. More insured value means more demand for cat bonds

AI data centers are incredibly expensive. Private data center construction jumped a staggering 85% from December 2023 to December 2025. According to McKinsey, data centers will require nearly $7 trillion dollars to keep pace with demand by 2030. Of course, as thousands of new facilities are built, insurers must cover much larger concentrations of property risk than before.

Traditional insurers and reinsurers cannot typically absorb all of this exposure on their own, so we could expect them to transfer some of the risk to capital markets through cat bonds and other insurance-linked securities.

  • AI data centers create "accumulation risk"

While the amount of AI construction is going up, it’s also being concentrated in certain areas because of access to power, land, and tax incentives, such as the south and MidWest U.S. with the largest number of facilities in Virginia and Texas.

If a major hurricane, tornado outbreak, flood, wildfire, or power-grid event hits one of these clusters, losses could be highly-correlated across many facilities at once. For cat bond investors, this could mean larger potential loss events, greater correlation among insured assets, and the need for more sophisticated modeling for natural disaster events.

  • Higher Yields & Greater Diversification

This item is directly related to how more insured value means more cat bond issuance. The AI infrastructure boom has the potential to create a new asset class within the insurance-linked securities market because insurers and brokers could explore cat bonds specifically tied to data-center risks, including hurricanes, floods, earthquakes, fires, or cyber attacks. As a result, overall supply of cat bonds might increase, and these cat bonds might also carry higher yields if the risks are difficult to model. All of this might create potential opportunities for more diversification for cat bond investors.

  • Cat-bond pricing may become more attractive

If AI data centers increase aggregate catastrophe exposure, insurers will seek more protection, which can push cat-bond spreads higher. So as demand goes up and insurers have higher premiums, this could mean greater returns for cat bonds investors. In fact, the market is already seeing interest in bringing alternative capital, such as hedge funds, pension funds, and ILS funds, into data-center risk transfer.

  • Secondary effects on existing cat bonds

AI data center construction can also affect cat bonds that were not originally designed around AI. For example, a hurricane striking a region with a large concentration of new data centers may generate larger insured losses than historical models anticipated, which could increase expected losses for regional property catastrophe bonds and force catastrophe models to be updated to reflect the rapid growth in insured values around data-center hubs.

At a high-level, the AI data-center buildout is generally bullish for cat bond issuance volume because it creates large, new insurance needs that traditional reinsurance capacity may not be able to satisfy. However, it may also increase catastrophe risk concentrations, especially where AI facilities cluster in disaster-prone regions, potentially leading to higher yields, more issuance, and more complex risk modeling for cat-bond investors.

As the AI boom continues, investors will need to consider potentially higher yields to the possibility of large, correlated losses from natural disasters affecting major data-center clusters.

 

Sources:

McKinsey

The Motley Fool

Pew Research

Moody’s