Data Centers: Key Trends and Insights for Owners and Occupants
This article originally appeared in the Q2 2025 edition of The Evaluator.
Data centers are critical infrastructure in today’s world, but they present unique and significant challenges for property tax valuation. Many jurisdictions have provided property tax incentives or abatements for newly built facilities, but will this trend continue? What happens after the property tax incentive ends? Demand for data centers is anticipated to grow rapidly—driven by cloud computing, AI, and digital services. However, many tax assessors lack the expertise to value data centers correctly, potentially leading to inaccurate assessments or over-assessments, causing owners and tenants to pay too much in real property tax.
Overview
As of March 2025, the United States leads the world in the number of data centers, with approximately 5,426 operational facilities. As demand continues to increase, there has been a shift from large markets due to factors such as energy capacity, availability of real estate, favorable zoning regulations, and costs. The availability of affordable energy and tax incentives has made new markets attractive for hyperscale and colocation facilities.
The Valuation Paradox
When it comes to real property tax assessments, data centers are often misunderstood and often overtaxed. What drives these assessments? Large data center developments often involve significant investment, resulting in complex financing deals and announcements of substantial capital investment. However, tax assessors struggle to value data centers accurately. While there is often significant capital investment, the majority of the investment in a data center is in the business components, which are not real estate and therefore should not be taxed as such.
Property taxes are often the largest single operating expense (assuming the property is not receiving property tax abatements). While there may be up-front incentives abating property taxes for an initial period, what happens after the abatement expires?
The Valuation Challenge
The real estate component of a data center is similar to a warehouse. However, a sophisticated data center will have many specific requirements that far exceed the requirements for a typical warehouse. These requirements include substantial electrical components to operate the servers, with the ability to have an uninterrupted power supply and redundancy to ensure the facility operates 24/7/365. Additionally, the data center must have structural improvements to support the fiber optics and cabling, as well as reinforced or specialized flooring to support the weight of equipment, battery backups, and cooling systems. The facilities also often have redundant electrical systems, extensive water or air handling components, double roofs, and increased security requirements. All of these components are constructed with the facility but, in most cases, may not meet the definition of real estate for taxation purposes in a particular jurisdiction. Real estate taxation should only be based upon the value of the real estate. However, because modern data centers are still relatively new in most jurisdictions, many assessors haven’t determined how to exclusively value only the real estate without incorporating the value of the non-real estate components that are often difficult to separate from the real estate itself. Owners and operators should not lose sight of the fact that even if the property is receiving tax incentives, community development charges could be tied to valuation, or the incentive will end or ratchet down over time, exposing the owner or operator to a tax burden not tied only to the value of the real property.
Valuation Methods Applied to Assess Data Centers
As with any real estate, tax assessors employ three standard methodologies to value the real property: the cost approach, the sales comparison approach, and the income approach to value. Which approach is the best to value only the real estate for a data center?
All three approaches can be useful; however, all have potential issues that must be evaluated to determine the proper assessment of the real property. Let’s first consider some of the flaws with the different approaches to value. Both the sales comparison approach and income approach may not be appropriate for the valuation of a data center.
Starting with the sales comparison approach, a data center is likely to trade in the market as a business operation or going concern. This means that the buyer is purchasing more than just the real estate. The buyer is purchasing the personal property and potentially service contracts or other intangibles. Further, the purchase of a build-to-suit property for a specific user is going to include not only the real estate but also the personal property or business fixtures that should not be considered real property. Any knowledgeable market participant understands that the purchase of a data center is more than just the building and land, making sales of data center facilities suspect to value only the real estate.
Second, the income approach to value suffers from some of the same issues as the sales comparison approach – the inability to isolate only the value of the real estate. There are numerous types of leases and service agreements in the data center industry. The more equipment and services included in the lease agreement, typically leads to a higher lease rate. But again, this is not only for the use of the land and building but often includes components that are considered personal property or business fixtures in certain jurisdictions or services such as security. Leases can vary substantially from one to another. With a NNN lease, the tenant would rent shell space in the data center and be responsible for connecting to power, its equipment, cooling, and security. At the other end of the spectrum would be a lease that could provide full service, including equipment and utilities. These two different lease structures demonstrate the difficulty in evaluating the components included within the lease rate and, in part, account for the differences in the lease rate itself. But they do not effectively account for the rent associated with the real estate alone. Instead, the exercise of trying to isolate the contributory real estate component of the overall lease rate is speculative at best, especially as other factors such as access to energy and the necessary infrastructure (powered shell) may be increasingly driving rental rates for data centers.
Finally, the cost approach to value data centers can be a useful method if properly developed and if only items that are real estate are included within the approach to value. If the cost approach is properly developed, it will best reflect the value of the real estate. A cost valuation approach starts with estimating the cost, often through valuation manuals. From there, depreciation is determined, and the value of the land is added to the depreciated improvements. This results in a determination of the fee simple value for the real property. However, there are still challenges with this approach, especially in jurisdictions that do not tax personal property or have business fixture definitions that may exclude certain components from the definition of real property. With these factors in mind, the cost approach is the best method to isolate the value of only the real property.
Final Thoughts
As data centers become more prevalent across the country, the impact of real estate tax assessment will evolve and change. However, until each tax assessor recognizes the flaws with the methods they employ, data center owners and occupants should carefully review their real property tax assessments to determine if the value reflects only the value of the real property. Even if the properties are subject to tax incentives, the review process is still important because there may be other charges such as community development charges tied to value, and because, importantly, the property will become fully taxable in the future.
[View source.]
