Finding Power and Optimal Sites for Data Centers on a Constrained Grid

The rapid expansion of AI is accelerating data center development across the United States, but access to reliable power is becoming a defining constraint. In this first article, we examine how market intelligence can help developers identify viable sites, understand grid conditions, and make earlier, more confident siting decisions.

Data centers have become a major growth market in the United States. Hyperscalers such as Amazon Web Services (AWS), Microsoft Azure, and Google Cloud are investing at record levels, fueled largely by the rapid expansion of AI.

As facilities grow in size and number, their electricity demand is expected to surge, putting added strain on energy resources and grid infrastructure. According to some industry estimates, U.S. data center electricity demand could increase fivefold by 2035, reaching 176 gigawatts (GW). Research also shows that access to power remains the top consideration when selecting greenfield development sites.

These constraints are forcing developers to evaluate power availability much earlier in the process. Too often, grid capacity is assessed only after critical siting, permitting, and investment decisions have been made, increasing risk and limiting options for large-scale projects. To make smarter decisions sooner, developers need a clearer view of market conditions, grid dynamics, and interconnection realities.

Solutions such as those from Hitachi Energy’s Energy Portfolio Management group help bring these complex market signals together, giving data center developers a more complete and actionable view of the U.S. market.

Effective market intelligence begins with a detailed understanding of each facility in the data center landscape. At the most basic level, this includes core identifiers such as site name and location. More advanced insights, often derived from primary research, add critical context such as operator, load capacity where available, and development status, including whether a site is existing, under construction, or proposed (see Figure 1).

Additional attributes help developers understand how facilities are intended to operate and how they may affect local power demand. These include the data center’s primary use and IT processing profile, such as colocation, crypto mining, enterprise, edge, hyperscale, or exascale. Market intelligence solutions can also reveal whether a site is carrier-neutral, the type of property it occupies—such as standalone, mixed-use commercial, colocation tenant, build-to-suit, powered shell, or underground—and whether it functions as a cable landing station.

As data center development accelerates, it is increasingly important to understand how new facilities connect to the power grid, how much load they are expected to add, and what those connections mean for capacity at specific locations. Figure 1 illustrates the proximity between data centers (represented by colored shapes), nearby substations (depicted as white circles), and transmission lines (shown as colored lines indicating specific voltages), helping developers visualize potential interconnection pathways and evaluate local grid conditions earlier in the siting process.

Visual proximity is only the starting point. A deeper level of analysis comes from confirming relationships between data centers and substations, including the associated buses, to better understand available withdrawal capacity at a given point of interconnection (POI). This “Transmission Bus Withdrawal Capacity” helps developers assess how much power may be available from the grid, whether they are screening large territories across multiple sites, as shown in Figure 2, or evaluating specific substations and POIs in greater detail, as illustrated in Figure 3.

Once relationships between data centers and substations are established, analysts can extend the assessment to the transmission lines that serve those connection points. This is critical because congestion and outages on nearby lines can materially affect both power prices and reliability. Figure 4 shows how visualizing flowgates (yellow highlight on transmission line in Figure 4) alongside binding constraint data and locational marginal price (LMP) patterns can reveal where grid stress may create operational or financial risk.

Another valuable metric is the Distribution Factor, also known as the Shift Factor, which helps identify where congestion charges are most likely to arise and which grid elements are driving those constraints. By linking these signals back to infrastructure near prospective data center sites, developers can make more informed decisions about siting, interconnection strategy, and long-term risk.

As data center load grows, developers must also understand the generation resources that could support that demand. This requires looking beyond the load side of the grid to identify proposed generation projects near prospective data center sites, along with key attributes such as ownership, capacity, fuel type, and development milestones (see Figure 5). Just as importantly, developers need visibility into where those generators plan to interconnect, their position in the interconnection queue, and how much injection capacity may be available at the relevant bus or substation. Together, these insights help clarify whether nearby generation can realistically support future data center development and on what timeline.

Broader market indicators add another important layer to data center analysis. Generation mix data helps developers and operators understand the composition of local power supply, including the availability of cleaner energy resources and the operating patterns of nearby generation assets (see Figure 6). By examining supply sources, output trends, capacity factors, and load dynamics, stakeholders can gain a clearer picture of how specific power markets function and what those conditions may mean for long-term development decisions.

Map layers that outline electric service territories and gas utility area footprints are also helpful in understanding the market dynamics for potential data centers and generation units. Additionally, the ability to gain early insight into political sentiment and regulatory risk enables proactive stakeholder engagement and more confident site selection ahead of public disclosure. 

Power purchase agreement (PPA) intelligence adds another valuable perspective. Sources such as the Federal Energy Regulatory Commission’s Electric Quarterly Reports (EQR) can provide visibility into announced pricing between energy buyers and sellers, including renewable project owners and data center operators. That information can support price benchmarking, reveal market activity, and strengthen competitive intelligence during procurement and contracting discussions.

Natural gas infrastructure is also becoming a more important consideration in data center siting, particularly as developers evaluate a broader range of power supply options. Understanding pipeline utilization rates and capacity constraints, key flow points and daily scheduled flows, customer mix, and operational notices can provide a clearer view of regional natural gas market conditions. Trends in natural gas pricing and storage can also materially affect project economics and long-term operating risk.

Whether a company is injecting power as a generator or withdrawing power directly from, the grid as a data center operator, identifying locations with favorable pricing remains essential. Historical pricing analyses—whether monthly, annual, hourly, peak-period, congestion-specific, or spread-based—can reveal patterns that materially affect both site selection and operating strategy. Paired with locational marginal price visualizations such as the one shown in Figure 4, these insights help developers and operators identify locations with more favorable economics and better-informed risk profiles.

As this first blog shows, securing power for data center growth begins with a connected view of the market. By bringing together siting insights, grid relationships, generation data, and broader market signals, developers can make better-informed decisions earlier in the process and reduce risk before major investments are made.

In Part 2, we explore how those same insights can support long-term planning, procurement, trading, and operational decisions. Together, the two articles show how Hitachi Energy helps data center customers move from identifying viable sites to managing power more strategically across the full project lifecycle.