Rob McGibney, executive vice president at KB Home, told analysts on the company's most recent earnings call that several community openings had been delayed by an obstacle that no amount of framing speed or permit acceleration could fix. "We've got a number of communities that we know are coming, we just don't know when," he said. "They're generally ready to go outside of waiting for transformers to be installed and energized."
Drywall hung, paint dry, cabinets in, countertops set, landscaping mulched.
And the house sits there, completely dark.
Not because of an inspection failure or a punch-list dispute or a buyer who changed their mind about the backsplash, but because the green metal box that steps 7,200 volts down to 240, the pad-mount distribution transformer that every subdivision needs before a single light switch works, hasn't arrived from the factory yet, and nobody at the utility can say when it will.
Eighty percent of builders and apartment developers are reporting transformer and electrical equipment shortages, according to the National Association of Home Builders, with delays in some markets stretching from 26 to 48 weeks. In Duvall County, Florida, one of the state's fastest-growing residential construction regions, the wait hit 30 weeks before Hurricane Ian compounded the backlog and pushed some projects past a full year. Arkansas and the Gulf Coast have exceeded a year.
Where the Transformers Went
There are between 60 million and 80 million distribution transformers in service across the United States, according to the National Renewable Energy Laboratory. More than 55 percent of them are over 33 years old, approaching or exceeding their designed lifespans and generating replacement demand on top of new construction requirements.
Then data centers arrived. They brought checkbooks that make residential procurement look like a lemonade stand.
Demand for generator step-up transformers has increased 274 percent since 2019, according to a September 2025 report co-authored by Wood Mackenzie and American Clean Power. Substation transformer demand rose 116 percent over the same period, driven overwhelmingly by AI data center construction spending that climbed 96 percent in three years, alongside renewables projects and the broader electrification push that quietly doubled peak demand projections in regions that hadn't built a new substation since the Clinton administration. A single hyperscale facility can require 50 to 100 transformers, while a 50-home subdivision needs three to five. When both orders land at the same factory floor, the manufacturer does not need long to decide which one to run first.
Thomas Murphy, vice president at Power & Construction Group, an electrical-infrastructure contractor near Rochester, New York, put it bluntly to the Wall Street Journal in July 2026: "Anything that is available in the United States is gobbled up immediately. Between data centers and solar projects, they're taking a huge bite of any equipment."
The Numbers Behind the Wait
The North American Electric Reliability Corporation reports that lead times for large power transformers reached 210 weeks in 2024. Four years. Even smaller distribution transformers used for residential service are backordered as much as two years.
Prices have followed the lead times upward. Burns & McDonnell's Michael Novev told Reuters Events that transformer prices jumped roughly 80 percent over the last five years, a figure Murphy corroborated at 70 percent. Utilities report costs four to six times their pre-2022 levels. Switchgear? Forty-four weeks average lead time by Q2 2025, with circuit breakers up 47 percent since 2021 and medium-voltage switchgear up 50 percent. Every link in the electrical supply chain is strained, and every link is getting worse.
The supply side offers no relief, and this is the part that keeps procurement managers awake. The cores of most U.S. transformers require grain-oriented electrical steel, a specialty product with specific magnetic properties manufactured domestically only by Cleveland-Cliffs at two plants in Pennsylvania and Ohio, which means every domestic transformer manufacturer depends on a single supplier for its most critical input. Eighty percent of large transformers have historically been imported from Mexico, China, and Thailand, but tariffs have tightened that valve and forced domestic manufacturers to compete for raw materials they were never tooled to source at this volume. Section 232 duties of 50 percent on steel and aluminum remain in place even after the Supreme Court's February 2026 ruling in Learning Resources, Inc. v. Trump struck down broader IEEPA tariff authority. Some firms have begun purchasing factory production slots at a premium before they even have a project site or a customer under contract, simply to guarantee access to equipment they know will not be available on the open market six months from now, according to Burns & McDonnell. Speculative transformer hoarding is where we are.
The Carrying-Cost Math Nobody Is Doing
Here is where the project management breakdown becomes expensive.
A builder running a 50-home subdivision with a construction loan at 8.5 percent on an average home cost of $450,000 is paying roughly $105 per home per day in interest alone. Add insurance, property taxes, HOA compliance deadlines, staff overhead, and the superintendent you are paying to check on finished houses that nobody can move into, and the all-in carrying cost runs $140 to $180 per home per day. If the entire subdivision is finished but waiting four months for pad-mount transformers, that is 120 days of dead carrying cost across 50 homes, and at $160 per day the total reaches $960,000 across the subdivision.
Nearly a million dollars. Burned waiting for a metal box.
Shad Schmid, president of the Greater San Antonio Builders Association and co-founder of King Fish Development, told Builder that lead times in his market peaked at 12 to 18 months. In areas where rising interest rates cooled demand, the wait has compressed to roughly five months, but that is still three to four months longer than the historical norm of four to eight weeks.
CPS Energy in San Antonio confronted this directly after a 40-percent surge in residential transformer demand hit a three-year high and threatened to leave entire subdivisions sitting dark through the end of the fiscal year. Their solution was a multi-prong approach: diversified sourcing, inter-utility buying and selling, refurbishing existing units, and increased vendor engagement. It worked well enough to secure roughly 4,500 transformers for 2024, a figure that would have been unremarkable a decade earlier.
Most builders do not have a utility's purchasing power or the option to refurbish their own equipment. They wait.
The Fix Nobody Is Specifying
An AI-optimized smart electrical panel costs $3,000 to $5,000 per home, roughly ten times a standard load center. Builders look at that number and stop reading. Here is why they should keep going.
Smart panels from companies like Span, Lumin, and Schneider Electric's newer QO series use real-time AI load management to dynamically prioritize circuits. An EV charger, a heat pump, and an electric range do not all draw peak load simultaneously in a house with active load orchestration. The practical effect: the transformer serving that cluster of homes sees a measurably lower coincident peak demand. If every home in a 50-unit subdivision is running a smart panel, the utility can approve a smaller transformer bank or approve service with fewer units, because the aggregate peak the transformer must handle drops by 20 to 30 percent in modeled scenarios.
Run the numbers on a subdivision where specifying smart panels reduces the transformer requirement from five pad-mount units to three. Two fewer transformers to source, two fewer to wait for, and the months of carrying cost saved at $160/home/day dwarf the $200,000 investment in panels ($4,000 × 50 homes). Even if the reduction is only one transformer eliminated, the ROI closes within the first month of avoided delay, and the homeowner gets a panel that will manage their EV charging, solar export, and battery dispatch for the next 20 years.
Nobody is doing this calculation, because builders have never had a reason to think of transformer procurement as a variable they can influence rather than a fixed administrative step. They submit a utility service request, they wait, they build around whatever timeline the utility hands back, and that has worked for forty years because the timeline used to be six weeks. It is not six weeks anymore. The equipment shortage has turned a routine administrative step into the critical path, and the critical path demands a different kind of attention than any builder's project management playbook currently allocates to electrical interconnection.
The Grid Itself Is Short
PJM Interconnection, the regional transmission organization serving 67 million people across 13 states and the District of Columbia, announced in July 2026 that its capacity auction for the 2028/2029 delivery year came up 6,831 megawatts short of the reliability requirement. This was the second consecutive auction with a shortfall, and it marked the first time in PJM's history that the entire regional transmission organization fell below the standard designed to prevent blackouts. The clearing price hit the FERC-approved ceiling of $325 per megawatt-day.
That shortfall will show up in utility bills. It will also show up in utility service timelines, because when the grid operator itself cannot secure enough generation capacity to meet projected demand, the equipment that connects new homes to that grid does not become more available; it becomes more rationed.
The Strongest Counterargument
The transformer shortage predates the data center boom. COVID-era supply chain disruptions, the loss of domestic manufacturing capacity over decades, and chronic underinvestment in grid infrastructure all contributed to a system running without slack. If data centers vanished tomorrow, 55 percent of distribution transformers would still be aging past their design life, Cleveland-Cliffs would still be the sole domestic GOES steel producer, and Section 232 tariffs on imported steel and aluminum would still inflate costs.
This argument is structurally correct, and it deserves its full weight. Data centers did not create the fragility. But they converted a slowly deteriorating situation into an acute crisis by dropping 274 percent more demand onto a supply chain that was already failing to keep pace with routine replacement needs. Think of a bridge rated for 20-ton loads that needed repair for years while everybody involved knew it was deteriorating and filed reports about its condition that nobody acted on. The 60-ton truck is what caused the collapse. Nobody disputes the truck.
What This Article Did Not Prove
The carrying-cost scenario above uses representative figures. Construction loan rates, insurance obligations, and carrying-cost structures vary by market, lender, and project type. The smart-panel-to-transformer-reduction math is based on modeled coincident peak scenarios, not a published field study in which a utility formally approved fewer transformers for a subdivision specifically because of smart panel installation. That study does not exist yet, and it should.
Regional variation matters enormously: a builder in a high-growth Sun Belt market where CPS Energy is actively refurbishing transformers faces a different timeline than one in a rural Midwest county where the utility has no refurbishment program and a single domestic supplier. The 44-week switchgear lead time is a Q2 2025 average; individual orders may come in faster or far slower.
What the data does establish: the same AI boom that promised to make construction faster is consuming the physical equipment that construction requires to deliver a habitable home, and the residential sector lacks the purchasing power, the procurement sophistication, and the political attention to compete for what remains.
← Back to all articles