Meeting demand: Identifying energy risks with new data centers

Electricity accounts for 40% to 70% of the operating costs of data centers, which support the advanced computing systems on which societies increasingly depend. Innovations in artificial intelligence (AI) are fueling this demand. For example, a single ChatGPT query consumes nearly 10 times more energy than a traditional Google search. Globally, data center energy demand is projected to more than double by 2030.

Getting more renewables on stream

Whereas data centers have historically relied on fossil fuels for power, hyperscale data center developers are increasingly exploring alternative power generation options, including nuclear, hydrogen, solar, battery storage, and wind. The significant growth in renewable projects in recent years has helped position renewables as an attractive power source alongside natural gas.

However, there are concerns that delays in power projects could slow the development of new data centers. Currently, there is a backlog of renewables projects seeking interconnection across the US, with an average wait time of four years. Similarly in Europe, businesses are facing five to eight-year wait times to connect to the continent’s electricity grids. However, this process can be expedited by addressing procedural inefficiencies, as demonstrated in Texas, where the wait averages just 1.5 years.

Investors may be concerned about the availability and reliability of power for new data centers, as a stable power supply is essential for these projects. For instance, servers cannot be taken offline, and cooling systems must operate continually.

Many developers are working on ways to reduce the energy demands of data centers on grids. For example, while user queries are processed immediately, other tasks, such as training new AI models, could be paused during peak demand.

Developers are also increasingly adopting a “behind the meter” strategy, which involves building data centers alongside renewable, nuclear, or natural gas assets and using the grid as a backup. Natural gas remains one of the leading power sources, although advance planning is important as gas turbines require a long lead time to build.

Insurance products can also support the launch of new renewable energy projects. For example, surety bonds as an alternate to letters of credit can offer an efficient way to securitize interconnection agreements.

Power availability is not the only factor to consider when setting up new data centers. While AI capabilities may be new, data centers are not and have common risks.

Understand multiple risks

Constructing a new data center involves assessing various risks, including counterparty credit risks, natural catastrophe risks, construction delays, permitting regulations, and any compliance with government policies. Creative approaches to addressing risks can often lead to more effective strategies for purchasing insurance products designed to mitigate them.

For instance, construction-related insurance products, such as delay in startup (DSU) insurance, are increasingly required by lenders. Parametric products also offer an alternative for addressing weather risks, particularly there may be limitations on traditional insurance.

Additionally, tax insurance can be an important enabler of investments that rely on anticipated tax credits, such as those from the US Inflation Reduction Act. As data centers become operational, more common insurance products like property, casualty, cyber, and D&O policies are also important.

Technology performance insurance can result in greater confidence among lenders and investors for more novel technologies developed by companies whose balance sheets may not support long-term warranties. For example, investment decisions for data centers might assume an anticipated performance of solar panels or batteries over long periods with expected degradation curves. Insurance may be able to mitigate the risk of these units degrading earlier than expected or of the manufacturer failing to honor the warranty. Some energy and hyperscale tech firms are seeking innovative ways to achieve economies of scale by purchasing insurance that covers risks affecting multiple data center and power generation projects.

Identify and mitigate your specific risks

The rise in data centers requires early engagement and thorough risk assessments, as companies must carefully manage traditional and emerging risks. For instance, co-locating data centers with electricity generating facilities necessitates expertise in the specific risks associated with each power sub-sector.

Marsh’s Energy & Power team works closely with companies in this area to help with risk management and insurance purchasing strategies. To find out more, please contact your Marsh representative.

Article

Meeting demand: Identifying energy risks with new data centers

Getting more renewables on stream

Understand multiple risks

Identify and mitigate your specific risks

Our people

Michael Gaudet

Related insights

Wildfire liability: Proactive strategies for utilities and infrastructure

Risk in Context Podcast: Insurance’s critical role in a successful energy transition

Managing steel and aluminum tariff uncertainties

Critical trends in liquefied natural gas (LNG): What you need to know

Marsh.com Sign in

Meeting demand: Identifying energy risks with new data centers

Getting more renewables on stream

Understand multiple risks

Identify and mitigate your specific risks

Our people

Michael Gaudet

Related insights

Wildfire liability: Proactive strategies for utilities and infrastructure

Risk in Context Podcast: Insurance’s critical role in a successful energy transition

Managing steel and aluminum tariff uncertainties

Critical trends in liquefied natural gas (LNG): What you need to know