13 Nov 2026
Why Are Grid Transformers a Bottleneck for AI Data Centers?
Transformers are the electrical devices that raise and lower voltage as power moves from generators to transmission lines to local grids and finally into buildings, and every new data center, power plant and grid upgrade needs them. Demand has outrun factory capacity: Wood Mackenzie reported in 2025 that lead times for large power transformers had reached about 128 weeks, with a supply deficit of about 30%. Because a data center cannot open without its grid connection and electrical gear, transformers and related equipment have become one of the gating factors for the AI buildout.
What does a transformer do?
Electricity travels most efficiently over long distances at very high voltage, but it is generated and used at lower voltages. Transformers bridge those levels using coils of copper or aluminum wound around a core of specialized steel. Three broad types matter for data centers:
Generator step-up transformers raise the output of a power plant to transmission voltage.
Large power transformers sit in substations and connect the high-voltage transmission network to regional grids. Many weigh more than 100 tons and are custom built for each site.
Distribution and pad-mounted transformers lower voltage further for use inside campuses and buildings.
A large AI data center campus may require its own high-voltage substation with several large transformers, plus dozens or hundreds of smaller units inside the site. Our article on how data centers work shows where this equipment sits.
How long are lead times?
Wood Mackenzie’s 2025 analysis put average lead times at about 128 weeks for power transformers and about 144 weeks for generator step-up units, with U.S. supply falling about 30% short of demand.1 Before 2020, lead times for many units were measured in months rather than years. Switchgear, circuit breakers and high-voltage cable have seen extended lead times as well.
Why has supply fallen behind?
Demand from many sources at once. Data centers are only one driver. Utilities are replacing aging equipment, much of it installed decades ago. Renewable and storage projects each need step-up transformers. Electric vehicle charging, new factories and population growth in some regions add load.
Specialized materials. Transformer cores use grain-oriented electrical steel, produced by a small number of mills worldwide. Windings use large amounts of copper.
Custom engineering. Large transformers are often designed for a specific substation, which limits standardization and makes inventory difficult to hold.
Limited factories and skilled labor. Manufacturers reduced capacity during years of flat demand. New plants take years to build, and winding and testing large units requires experienced workers.
Import dependence. The United States imports a large share of its large power transformers, which exposes supply to trade policy and shipping.
How does this connect to grid interconnection?
Transformers are one part of a broader interconnection bottleneck. Lawrence Berkeley National Laboratory’s 2026 “Queued Up” report found about 2,061 gigawatts of generation and storage capacity waiting in U.S. interconnection queues, with a median of about 61 months from interconnection request to commercial operation for recently completed projects.2 New large loads such as data centers face their own study and upgrade processes. Even when power is available, the substation, transformers and transmission upgrades needed to deliver it can take years. Our article on AI power and electrification discusses the demand side.
What does a transformer shortage cost?
Delays carry real costs. A data center that is finished but cannot energize earns no revenue while its owner continues to pay for land, construction financing and idle equipment. Utilities facing equipment shortages may postpone replacing aging units, which can raise the risk of outages, and may hold scarce spares for emergencies rather than new connections. Prices have risen as well: buyers have reported paying substantially more per unit than before 2020, reflecting higher material costs and strong demand. These costs ultimately flow into project budgets, utility rates and the pace at which new computing capacity comes online.
How are developers responding?
Developers and utilities have adopted several strategies: ordering transformers years in advance, sometimes before sites are final; standardizing designs to allow reuse across projects; reserving factory slots through framework agreements; building on-site generation so a campus can operate before its full grid connection arrives; and buying refurbished or spare equipment. Manufacturers have announced expansions of U.S. and North American transformer factories, though new capacity takes time to come online. Our article on powering AI with gas, nuclear and SMRs covers on-site generation options.
What does this mean for investors?
Transformers illustrate how AI spending reaches into traditional industrial companies. The relevant supplier base includes electrical equipment manufacturers, electrical steel producers, copper producers, switchgear makers, engineering and construction firms and utilities. Many of these companies serve broad grid and industrial markets, so data center demand is one driver among several. Risks include cyclical demand, raw material costs, competition from new factories and the possibility that demand slows if data center plans change. Our article on traditional versus digital infrastructure explains how these categories overlap.
Frequently asked questions
Why can’t manufacturers just build more transformers?
They are expanding, but large transformers require specialized steel, copper, insulation and testing equipment, as well as skilled workers who take years to train. New factories take several years to plan, build and qualify. Manufacturers also weigh the risk of overbuilding after past periods of weak demand, which encourages measured rather than rapid expansion.
What is grain-oriented electrical steel?
Grain-oriented electrical steel is a specialty steel processed so that its crystal structure aligns in one direction, reducing energy losses when magnetic fields pass through it. It forms the core of most power transformers. Only a small number of mills worldwide produce it, which has made it one of several input constraints for transformer manufacturing.
Do data centers need their own substations?
Large data center campuses often do. A campus drawing hundreds of megawatts typically connects directly to high-voltage transmission lines through an on-site or nearby substation, which contains large power transformers, switchgear and relays. Smaller facilities may connect to existing utility substations at lower voltages, sometimes after the utility upgrades its equipment.
How long does it take to connect a data center to the grid?
It varies widely by region and project size. In some areas, large new loads have faced waits of several years for interconnection studies, substation construction and transmission upgrades. Transformer and switchgear lead times of two years or more add to that timeline. Some developers build on-site generation to begin operating before the full grid connection is complete.
Are transformer shortages only a U.S. problem?
No. Utilities and developers in Europe and elsewhere have reported extended lead times as electrification, renewable projects and grid replacement programs compete for the same factory capacity and materials. The United States is particularly affected because it imports a large share of its large power transformers and has fewer domestic factories for the biggest units.
How do thematic ETFs include grid equipment?
Electrification, infrastructure, utilities and some AI infrastructure ETFs may hold electrical equipment manufacturers, materials producers and grid contractors. In AI-focused funds, these holdings often sit alongside semiconductor and networking companies. Because many grid suppliers serve multiple end markets, reviewing their revenue mix shows how closely a fund’s holdings track data center spending.
- Wood Mackenzie, U.S. power transformer and distribution transformer supply analysis, second quarter 2025.
- Lawrence Berkeley National Laboratory, “Queued Up: 2026 Edition, Characteristics of Power Plants Seeking Transmission Interconnection,” 2026.
Cite this page: Adam Patti, "Why Are Grid Transformers a Bottleneck for AI Data Centers?," VistaShares, updated November 13, 2026, https://www.vistashares.com/grid-transformers-ai-bottleneck/. Plain-text version.
This material is provided for educational purposes only. It is not investment, legal or tax advice, and it is not a recommendation or an offer to buy or sell any security. Information is drawn from sources believed to be reliable as of the date of publication, is not guaranteed as to accuracy or completeness, and may change without notice. Figures attributed to third parties are their estimates. Projections and forecasts are not guarantees of future results. Investors should consult their own financial, legal and tax professionals before making investment decisions. Investments involve risk, including possible loss of principal. Companies in a single industry or theme may be affected by common economic, regulatory or technological developments and may be more volatile than the broad market. References to companies, governments or programs are for illustration only.