Developer alternatives if capacity is unavailable

If a site does not have room for interconnection today, the developer has other options to explore.

Transformer supply bottlenecks

It would seem straightforward to wait for a bigger transformer, but power transformer lead times, which ran seven to fourteen months before the pandemic, have now reached 160 weeks (3 years). Prices are up 80% or more in the past six years, and 80% of large power transformers used in the US are imported, as well as around 50% of distribution transformers. Domestic manufacturers currently meet only about a fifth of national demand. New domestic production capacity, with $2 billion committed across manufacturers like Hitachi and Siemens, is projected to come online by 2028. So that means ordering a new transformer is a multi-year capital commitment with a delivery date that's uncertain. Project financing is often sensitive to delayed starts and cannot tolerate the uncertainty and length of these wait times.

Energy storage

The developer can alternatively use batteries to shrink the interconnection request. Portland General Electric and GridCARE used hourly demand modeling and flexible resources, including batteries and on-site generation, to open up over 80 MW of data center capacity starting in 2026, with over 400 MW total by 2029. A conventional interconnection study would have taken years longer. Separately, Aligned Data Centers is installing a 31 MW, 62 MWh battery system in the Pacific Northwest to bring a site online faster than a traditional utility upgrade would allow, discharging during peak hours so the facility's peak draw from the grid stays below what would otherwise trigger a multiyear network upgrade.

Behind-the-meter

Beyond grid-connected solutions, the long and unpredictable wait for grid equipment upgrades is pushing data centers and other large loads to generate their own power. Behind-the-meter generation means building or contracting a generation asset on-site, consuming the power directly, and skipping routing it through the utility's interconnection process. One estimate puts behind-the-meter gas generation costs at $114 per megawatt-hour. This is 24% higher than the national average industrial grid rate of $92.

Developers plan to add 6.3 gigawatts of new natural gas-fired capacity in the US in 2026, driven largely by combined-cycle and combustion turbine units. Over 80% of this new natural gas capacity is concentrated in just five states, led by projects such as the Orange County Advanced Power Station in Texas and the Trumbull Energy Center in Ohio.

The inflated demand for gas because the stack that can be built fast enough ranges from fuel cells deployable in around 90 days up to heavy-duty turbines that still take three to seven years, and speed is the only criterion that matters once the alternative is a four-to-seven-year interconnection queue in many markets. There’s the chance that the interconnection timeline gets worse, at which point, the temporary solution could end up running for a decade or more.

A developer who learns in a week that a site has no nearby headroom still has the on-site generation option, the option of finding a different site with margin, or the option of accepting a multi-year wait with eyes open. But waiting a year and a half inside a formal interconnection study will lead to lost time on the alternatives.

Map radius

When searching a map to power a site, the search radius is a rough proxy for how much new distribution build a site can justify before it's cheaper to just look at a different substation instead. A site two miles from a substation with headroom and a site twenty miles from one are different projects entirely. One is a short tap-in versus miles of new feeder construction, and the point where that tradeoff stops making sense depends on how much load the project needs. A 50 MW data center can justify build out that a 2 MW facility probably couldn’t. The right way to set a radius depends on how much new construction the project would approve if the answer came back yes.