Power is now the first constraint checked, and in a growing number of regions, manufacturers are discovering that the capacity they assumed would be available has already been claimed by a data center campus down the road. The competition is not hypothetical. It shows up in interconnection queues, transformer order books, and utility large-load dockets across the country.
The scale of the mismatch is easiest to see in PJM, the regional grid operator spanning 13 states and Washington, D.C. Its most recent interconnection cycle drew more than 800 project applications totaling roughly 220 gigawatts (GW) of proposed generation, according to PJM Inside Lines. That volume of interest reflects a grid trying to catch up to demand growth that data centers and other large loads created, not a queue built with industrial expansion in mind. Getting into that queue is only the first hurdle. RMI, the clean energy research organization formerly known as Rocky Mountain Institute, found that the average timeline from interconnection application to commercial operation in PJM has stretched from under two years in 2008 to more than eight years by 2025, and that 74% of the 294 GW PJM has studied since 2020 withdrew before ever reaching service.
PJM's Reopened Queue Favors Whoever Applied First, Not Whoever Needs Power Most
PJM shifted in 2022 from a first-come, first-served model to a first-ready, first-served process, screening out speculative projects before they can occupy queue positions that industrial developers might otherwise use. The Electric Power Supply Association reports that its member generators have announced more than 12 GW of new capacity in PJM since mid-2024 through new builds, uprates, and delayed retirements, a response to the shortage that is real but slow relative to the pace of large-load additions. A manufacturer weighing a multi-year capital project now has to ask whether its timeline can survive a queue that was reformed for reliability, not for speed. In most cases, the honest answer is no.
The reform also introduced the Expedited Interconnection Track, a parallel pathway for large, advanced projects that can demonstrate full site control and post a readiness deposit of $15,000 per megawatt. PJM caps that track at ten approved projects per year, and it was built with large generation resources in mind, not industrial facilities trying to secure delivery capacity. For manufacturers, the practical takeaway is that the fast lane exists, but it was not built for them.
Utilities Don't Choose Winners. They Allocate Scarcity.
It is tempting to read all of this as data centers beating manufacturers to the punch, but that framing gives utilities more discretion than they actually have. A utility studying an interconnection request is not weighing whether a hyperscaler's mission matters more than a factory's. It is checking whether the applicant already has a signed load contract and a deposit on file, whether the engineering study has progressed far enough that construction could realistically start, and whether the local transmission system has room left to carry the additional draw. State commissions and FERC add their own sign-off on top of that. Data center developers tend to clear that bar faster, not because utilities favor them, but because they walk in with the offtake agreement signed, the deposit funded, and an engineering team that has run this exact process dozens of times before.
That distinction matters for how executives should read the competition. The system is not biased against industrial users. It is simply allocating a finite resource to whichever applicant proves readiness first, and readiness has become a function of capital and process discipline rather than industry type. A manufacturer that treats utility engagement as a late-stage formality will keep losing that race to whoever treats it as a parallel workstream from day one, regardless of what gets built next door.
Transformer Backlogs Are Turning Site Selection Into a Manufacturing Supply Chain Problem
Even a manufacturer with an approved interconnection agreement can be stalled by hardware. High-voltage transformer lead times that ran 24 to 30 months before 2020 now stretch as long as five years, driven largely by the same data center buildout competing for grid capacity. Eaton, Vertiv, and Schneider Electric have all announced new manufacturing capacity for grid equipment, but new factories take years to reach full output, meaning the shortage will outlast most current expansion plans. A facility that clears its interconnection study can still sit dark for a year or more waiting on the switchgear needed to actually accept power. The bottleneck has effectively migrated from the substation to the factory floor of the companies that build substations.
ERCOT and PJM Show the Same Squeeze Playing Out Differently
In Texas, the Electric Reliability Council of Texas (ERCOT) projects peak summer demand could reach approximately 145 GW by 2031, up from 85 GW in 2024, according to research compiled by Harvard's Belfer Center. Much of that growth traces back to the Dallas-Fort Worth data center corridor, drawn by competitive power prices and fast interconnection speeds that are now disappearing as the queue fills. Texas manufacturers considering expansion in the same corridor are effectively competing against the workload that made the region attractive in the first place. PJM's capacity market tells a parallel story: its most recent auction procured roughly 145,777 megawatts (MW), falling about 6.6 GW short of the reliability target and pushing capacity prices to a record. Both grids are telling the same story from different angles. Regional power markets are pricing in data center demand faster than they can build supply to match it, and every other large electricity user in that market pays the difference.
When the Queue Won't Move, Manufacturers Are Starting to Build Their Own Power
Facing years-long wait times, some industrial operators are adopting a strategy data center developers pioneered: generating power on site rather than waiting for the grid. The U.S. Department of Energy (DOE) issued emergency orders in January and May 2026 directing utilities to keep aging plants running through peak demand periods, and Energy Secretary Chris Wright has pointed to roughly 35 GW of underused backup generation nationally that could be called on during grid stress. That figure is a reminder that behind-the-meter capacity, long treated as an emergency reserve, is becoming a primary planning tool for any large power user willing to invest in it upfront. For manufacturers without a hyperscaler's balance sheet, that investment is a harder call, but it is increasingly the only way to control a timeline the grid no longer controls for them.
Site selection teams are responding by moving utility conversations earlier, sometimes before a location is finalized rather than after. A Kentucky data center campus that closed earlier this year came with its own substation and a path to 70 megawatts already secured, a structural advantage that manufacturers evaluating similar markets are starting to treat as table stakes rather than a nice-to-have. Copying data center tactics wholesale is not really the point, and few manufacturers have the balance sheet for it anyway. Power has moved from a procurement line item to a strategic asset that gets locked down before the rest of the project plan, not after.
