The slowest part of the buildout
If the AI buildout has a single hardest constraint, it may be the humble transformer. Transformers step voltage up for long-distance transmission and back down for use at the destination, and every data center, every substation, and every grid upgrade needs them. They are also among the slowest pieces of electrical equipment to build, and that combination of universal need and slow supply is why lead times have stretched to years. Adam O’Dell’s “Tech-Opoly” thesis leans on this scarcity, and it is one of the most defensible parts of the pitch.
Why transformers take so long to build
A large power transformer is not a mass-produced part. It is a custom-built, multi-ton machine wound with copper, filled with insulating oil, and engineered to a specific utility’s specification. Each unit takes months to manufacture, involves specialized labor and scarce materials, and must pass rigorous testing before it ships. There are only a handful of facilities in the world that can build the largest units, and they have been running at capacity for years.
The demand side has changed the equation. For decades, utilities replaced transformers on a steady schedule, and manufacturers sized their capacity to that slow drumbeat. Then data centers arrived and began ordering power in bulk, and the entire queue backed up at once. Our data center electrical equipment explainer places transformers in the full chain of scarce gear.
What years-long lead times mean for the sector
Lead times measured in years, and in some cases three or more years for the largest units, are the economic engine of the transformer story. They mean the manufacturers carry deep, multi-year backlogs, which in turn means revenue visibility that stretches well into the future and the ability to price each new order higher than the last. That pricing power is the difference between a scarce-product maker and a reseller, and it is why the theme has been so rewarding for the equipment names.
The same lead times also explain the risk. When backlogs stretch for years, the market starts pricing in earnings that have not happened yet, and the stocks that own those backlogs tend to run ahead of the fundamentals. Our grid infrastructure stocks explainer covers the “priced-in” question in detail.
Who builds the transformers
The transformer market is dominated by large diversified electrical manufacturers, the same giants that build switchgear and other equipment, with a smaller group of specialists around them. Eaton and Siemens are the household names, operating at massive scale across the electrical layer. The specialist makers focus on the industrial and utility-grade gear where the demand is most acute. The distinction that matters for investors is between a diversified giant, where transformer demand is a slice of a huge business, and a specialist, where the same demand is a much larger share of earnings.
The honest read
Transformers are a genuine, multi-year bottleneck, and the demand drivers, data center load growth plus a long-overdue grid refresh, are real and durable. The equipment names are well-positioned. The caution, as with the rest of the “Tech-Opoly” theme, is entry price, since the strongest stocks have already been re-rated higher. For the demand side of the story, our data center power demand explainer lays out why the load growth is real and where it is concentrated.
The materials and labor pinch
The transformer shortage is not just a demand problem; it is a supply problem on two fronts. The first is materials. Large transformers are built around a magnetic core made of grain-oriented electrical steel, a specialized product with only a handful of global producers, and around substantial quantities of copper and insulating oil. When transformer demand surged, those inputs tightened, adding cost and lead time on top of the factory bottleneck. The second is labor. Winding a large power transformer is skilled, physical work done by technicians who take years to train, and the pool of them has been shrinking for decades as the industry consolidated and offshored.
Neither of those constraints can be fixed quickly. A new electrical steel plant or a new cohort of trained winders takes years to materialize, which is why lead times for the largest units have pushed out to three or more years in some cases and why the shortage is expected to persist. For investors, that persistence is the bullish case, it keeps the backlog deep and the pricing power intact. It is also the risk, because a constraint this durable has already been priced into the stocks that own it.
NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.