Bernstein has laid out a framework for understanding how U.S. data center markets differ and why location matters for operators and investors. The firm divides markets into three types: Major Metro markets, Minor Metro markets, and Rural markets, each serving distinct workload profiles and facing different development dynamics.
Major Metro markets encompass what Bernstein describes as the top eight to 10 U.S. markets. These hubs are the primary destinations for latency-sensitive applications - the firm specifically names inferencing as an example - and for enterprise colocation. Analysts note these markets are difficult places to expand because they are already established and face long power queues, elevated construction costs, and constrained supply chains. Bernstein highlights that community resistance is on the rise - termed NIMBYism - particularly in Northern Virginia and Phoenix, driven in part by rising power costs.
Despite those challenges, the report singles out public data center owners Digital Realty Trust (NYSE: DLR) and Equinix Inc. (NASDAQ: EQIX) as concentrated in Major and Minor Metro markets and as being "among the most insulated from market noise." Bernstein argues that incumbent builders with established teams are positioned to manage rising local opposition and the operational complexities of these markets.
Minor Metro markets are defined as Tier 2 to Tier 4 cities or locations on the extreme outskirts of Major Metros. Bernstein frames these markets as the natural overflow destinations when capacity constraints or price points limit activity in Major Metros. They tend to attract workloads that are marginally less sensitive to latency or marginally more sensitive to price. The analysts note meaningful growth in these markets over the past five years and expect that trajectory to continue.
Rural markets, the third category, host workloads that are either latency insensitive or very price sensitive. Bernstein says these areas frequently feature larger facilities, a higher share of speculative construction, and developers who are often less experienced than those active in Major Metros. The advantages for rural sites include quicker access to power and lower build costs compared with dense urban hubs.
Bernstein uses Northern Virginia as a detailed example of Major Metro dynamics. The NoVa market is described as the largest U.S. data center cluster by installed capacity, with roughly 8 GW of supply and about 5 GW currently under construction. Power queues in the region are noted to be in the seven-year range. Pricing in NoVa is cited at about $217 per KW per month, and vacancy there remains under 1%.
The analysts also point to activity around Atlanta, labeling it another top U.S. market that is expanding rapidly. Quoted figures for Atlanta show pricing near $175 per KW per month and vacancy around 2%. Bernstein notes that, this quarter, both DLR and EQIX discussed large hyperscale-oriented developments outside Atlanta.
Overall, Bernstein expresses support for the strategy pursued by DLR and EQIX, calling market dynamics favorable and emphasizing that both firms are experienced builders with capable teams in Major Metros and face relatively limited execution risk. The firm carries Outperform ratings on both stocks, with price targets of $232 for DLR and $1,222 for EQIX.
Bernstein's note comes amid broader investor concerns about power delivery delays, local community opposition, and supply chain constraints that can affect the pace and cost of data center expansion. The firm’s segmentation of markets clarifies where those pressures are most acute and where operators may find the best returns for latency-sensitive and enterprise colocations versus price-driven workloads.