The two main factors that decide data center site selection are how long it will take the power grid to deliver power and whether it can deliver enough power. These variables determine the appropriate locations before clean power requirements are even considered.
While a data center can often be built in one to three years, new grid infrastructure can take five to 15 years to get through planning, permitting and construction. Even if a project gets connected to the grid, local congestion can limit what’s actually available on-site.
An interconnection queue is a waitlist of projects seeking to connect to a power grid. Electric grid operators manage these lists and often require impact studies before large-load projects, power generation or storage facilities can connect to the transmission grid.
According to Berkeley Lab data, there were about 8,200 projects waiting in interconnection queues at the end of 2025. For projects built in 2025, the average wait time between an interconnection request and operation was over five years.
While Berkeley Lab’s data focused on power generation and storage projects, it demonstrates how congested the connection system is and can help data centers understand the potential delays large-load interconnection projects face.
Just because there seems to be enough power in the region doesn’t mean enough power can actually be delivered to any specific site. Congestion is local, and data centers often cluster in the same places.
The International Energy Agency (IEA) reports that almost half of all U.S. data centers are located in just five regional clusters. Additionally, the IEA estimates that data center power constraints could delay about 20% of planned projects.
Whatever a building’s size, it can only draw what the local utility service can deliver. If the utility can’t handle the electrical load, the strain can damage the building’s infrastructure and equipment, making utility capacity and load analysis essential for data centers.
Under a September 2026 Executive Order, Massachusetts data centers above 25 MW must pay for their own grid upgrades and procure new clean energy to match their annual consumption or pay into a state fund that lowers energy costs for the public. Energy generation doesn’t have to be on-site, and the details of what counts as new still need to be defined.
Electric distribution companies must also take steps to remove large speculative data centers from interconnection queues. They should ensure that any new or expansion projects are legitimate to protect study resources and potentially decrease grid connection delays for other projects.
The clean power requirement is about where electricity comes from, not whether the grid can deliver it. A data center project can satisfy one requirement and still face issues from the other. Meeting the clean power requirement does not guarantee a site can receive enough power on the necessary timeline.
Grid headroom is local and time-dependent, so it varies between regions and even within them. Comparing markets involves comparing documented conditions, not scores.
Europe’s major hubs are the clearest documented examples of saturation. For example, Dublin and Amsterdam both have a high concentration of data centers and have paused new projects due to limited grid capacity. Developers here and in other concentrated locations, such as London and Frankfurt, face waiting in data center interconnection queues for 10 years or longer.
In the U.S., Texas has about 474 GW of requests, mostly from data centers, against a record peak of about 85.5 GW. The governor ordered an audit of all data centers in the Electric Reliability Council of Texas (ERCOT) interconnection queue. In response, ERCOT delayed its Batch Zero study, which covers the first projects under a new interconnection process for large loads.
Expensive delays in core markets are pushing investors to look beyond saturated hubs toward places with more available capacity. The IEA recommends that developers prioritize locations with spare capacity and consider on-site power options, which could help meet clean energy requirements and avoid data center grid connection delays. It highlights Spain and Finland as emerging hubs in Europe.
Developers should compare spare grid capacity, upgrade timelines, local queue volume and substation proximity when evaluating potential sites. Considering on-site power could reduce grid dependence and meet any clean energy requirements.
Anticipating data center growth, Spain has proposed tying grid access to renewable energy matching for new data centers. If implemented, this proposal would require new data centers with 1 MW or more of capacity to source at least 80% of their electricity from renewable sources, either through direct consumption or power purchase agreements.
The additionality requirement means every new megawatt consumed must come from a renewable source installed in the last 18 months. The goal is that installing new renewable energy sources will help keep electricity costs from increasing for consumers.
While this is an example of where policy could tie data center growth to clean power expansion, it doesn’t mean clean power rules are replacing grid capacity as the main site-selection constraint. Even with clean power targets, projects still depend on whether the local grid can physically accommodate the load.
While some places are tying grid access to clean power rules, grid capacity is still more likely to determine data center locations, since the grid ultimately reveals whether and when power can be delivered to the site. Clean power rules only dictate where the electricity comes from.
Developers should watch for Massachusetts’ implementation details and keep an eye on whether proposals like Spain’s become final. In the meantime, they can check grid capacity and connection timelines in proposed markets and consider secondary locations that may already have spare capacity.
