Transformer and Cooling System Makers Cash In on the AI Boom

Nvidia and other chipmakers have grabbed headlines for months thanks to the AI boom, but it's increasingly clear that the wave of orders runs much deeper into the supply chain. Makers of transformers, power systems and cooling equipment for data centers are seeing order growth in the tens of percent, and in some cases multiple times over.

A Bottleneck Beyond Chips

For more than a year, the narrative around the AI investment boom has centered on semiconductor makers - Nvidia, AMD and TSMC. But according to a Reuters analysis cited by wnp.pl, the real constraint on how fast data centers can be built is no longer the chips themselves, but power equipment: transformers, generators and cooling systems.

Hyperscale data center operators would like to bring new facilities online within six months of an investment decision. In practice, power infrastructure doesn't allow for that, since building and delivering it takes far longer than installing server racks.

Outside the industry, people talk about AI chips, but within the industry itself, people are certainly asking about lead times for generators and transformers - Wing Kin Cheung, CEO of BodaData

Record Order Backlogs

This shows up most clearly in transformer suppliers' results. South Korea's HD Hyundai Electric reported an order backlog worth $8.5 billion at the end of June 2026, up 23 percent from a year earlier. The company has already locked in a significant share of its production capacity for power equipment for the next three years.

We currently have an order backlog spanning more than three years, and a significant portion of our production capacity for major power equipment is secured for the next three years - HD Hyundai Electric

The company is also in talks with key clients about deliveries planned as far out as 2030, showing that data center operators are trying to secure power equipment years in advance rather than ordering on an ad hoc basis.

Chinese Maker Sees an Even Bigger Jump

A similar dynamic is visible at Chinese transformer maker Hainan Jinpan Smart Technology. The company reported that new data-center-related orders grew more than fourfold year over year in the first half of 2026, while its order backlog tripled. Demand is being driven by both domestic and foreign data center operators expanding capacity for AI workloads.

The geographic markets under the most pressure include North America, parts of Europe - including Finland, Germany and the United Kingdom - and the Middle East, where large investments in new data centers backed by local governments and funds are underway at the same time.

The Scale of Data Center Investment

The scale of the problem stems directly from the pace of investment. According to McKinsey estimates cited in the article, global data center spending is projected to reach nearly $7 trillion by 2030. That translates into electricity demand that transmission and distribution grids in many markets simply cannot keep up with at their current pace.

Delays in connecting new facilities to the power grid currently run as long as 24 months in emerging markets and over 8 years in developed markets, where regulatory procedures and transmission infrastructure upgrades take far longer. That makes access to transformers, generators and cooling systems a real factor in deciding who manages to launch a data center ahead of the competition.

What It Means for the Market

For investors tracking AI-boom-related companies, the signal is clear: exposure to this trend is no longer limited to chipmakers and cloud providers. Industrial power and cooling companies - previously seen as a stable but unglamorous business - are becoming beneficiaries of the same wave of demand, often facing less competition than in semiconductor manufacturing.

For data center operators, this in turn means having to plan power equipment purchases many months, and sometimes years, in advance, further stretching already tight construction timelines for new computing facilities to support AI models.