AI data center expansion hits supply chain bottlenecks in power and cooling systems

Updated on:04:34 Sep 2, 2026
Share:
  • Growing demand for transformers and cooling equipment is creating new bottlenecks in the AI data center boom
  • Innovation in solid-state transformers and liquid cooling technologies aims to improve efficiency and reduce environmental impact
  • Supply chain constraints and fierce competition threaten the growth prospects of key infrastructure suppliers

The AI race is creating an unexpected bottleneck, as power and cooling systems struggle to keep pace with increasingly demanding servers.

The global data center construction boom is generating profits not only for chipmakers like Nvidia. Demand is surging for transformers, generators, power distribution systems and cooling equipment as operators race to keep up with the development of artificial intelligence.

As stated by Reuters

According to McKinsey, global investment in data centers could reach nearly $7 trillion by 2030. At the same time, connecting new facilities to the power grid is becoming more difficult: in emerging markets, delays can last up to 24 months, while in major developed countries they can exceed eight years.

Hyperscalers, or the largest cloud infrastructure operators, often expect data centers to be completed within six months. As a result, power supply and cooling equipment have become critical parts of the supply chain.

Transformers convert high voltage from the power grid into levels suitable for servers, cooling systems and distribution equipment.

Transformer demand is growing alongside AI development

Leading transformer manufacturers, including South Korea’s HD Hyundai Electric and China’s Hainan Jinpan Smart Technology, recorded a sharp increase in orders related to AI infrastructure projects in the first half of 2026, particularly in North America.

HD Hyundai Electric said that demand in Europe is strengthening as US hyperscalers expand their investments in Finland, Germany and the UK. Interest in this equipment also remains high in the Middle East.

Its order backlog at the end of June had grown by 23 percent over six months to $8.5 billion. Data centers are expected to account for 16 percent of new orders in the company’s power equipment business next year, compared with 6.3 percent this year.

At Jinpan, the number of new orders from data centers more than quadrupled in the first half of the year compared with the same period in 2025. The related order backlog nearly tripled.

More powerful artificial intelligence chips consume increasing amounts of electricity, prompting equipment manufacturers to develop solutions that can improve efficiency and reduce the environmental impact of data centers. Water and electricity consumption are particular concerns.

According to Bank of America, the power consumption of a single AI server rack could exceed 1.5MW by the end of 2030. That is nearly 100 times the consumption of a conventional server rack.

One promising technology is the solid-state transformer. Unlike traditional devices, it uses semiconductors instead of large magnetic coils and copper windings to convert and distribute electricity.

UBS forecasts that such transformers could improve energy efficiency by approximately 4 percent and reduce costs. Mass commercial adoption is still at an early stage, but the technology’s share could rise to 40 percent by 2030. Chinese manufacturers are expected to strengthen their position thanks to their technological expertise and lower production costs.

HD Hyundai Electric and Jinpan are expanding their development of solid-state transformers. Taiwan’s Delta Electronics, one of the major suppliers of power infrastructure, has already reported using such devices in a small data center.

The data center cooling market is gaining momentum

Cooling systems are becoming another major area of growth. Operators are looking for ways to remove the heat generated by powerful artificial intelligence chips.

Bank of America forecasts that liquid cooling will be used in 70 percent of new AI data centers by 2030. Its current share is approximately 30 percent, compared with air-based systems. McKinsey estimates that liquid cooling can reduce electricity consumption by more than 27 percent.

Developers are also exploring unconventional infrastructure formats – including floating complexes, underwater data centers and servers housed in caves and tunnels. This is expanding the pool of companies that can profit from the industry’s development.

In particular, HD Hyundai Electric links the spread of autonomous power generation for data centers and floating facilities to new opportunities for suppliers of medium-speed marine engines.

Strong demand for thermal management equipment is supporting Delta Electronics, Taiwan’s Asia Vital Components and Auras Technology, as well as China’s Shenzhen Envicool Technology. All of them are part of Nvidia’s supplier ecosystem.

Competition and supply chain constraints

Despite the increase in orders, the rise in shares of power and cooling equipment manufacturers has slowed. Investors are assessing the risks of inflated company valuations, intensifying competition, project implementation delays and component shortages.

Delta Electronics shares have risen more than 90 percent since the beginning of 2026. Meanwhile, HD Hyundai Electric’s stock has remained almost unchanged after rising more than 100 percent last year.

China’s Jinpan and Envicool have lost nearly 30 and 20 percent, respectively, following sharp increases in 2025 of 118 and 244 percent.

Analysts advise bearing in mind that large-scale demand for equipment does not guarantee success for every supplier. The greatest advantages will go to companies that can attract key customers and ensure stable deliveries amid limited production capacity.

Thus, the trillion-dollar data center market is creating opportunities far beyond AI chip manufacturing. However, companies’ future growth will depend not only on demand but also on their ability to overcome equipment shortages, technological challenges and increasingly fierce competition.

Takeaways / FAQ

  • - What is the main bottleneck in the AI data center boom? Power delivery and cooling systems, not just chips.
  • - Why are transformers so important? They help convert grid power into levels usable by servers and cooling equipment.
  • - What is one major trend in cooling? Liquid cooling is gaining momentum as AI data centers become more power-intensive.
  • - Who may benefit most? Companies that can secure customers, deliver reliably and handle supply chain constraints.

Disclaimer: This article may have been created with AI assistance and reviewed by our editorial team. It is provided for general informational purposes only. Readers should verify information independently before relying on this content.

Subscribe Via RSS or Just Sign Up for Regular Updates
https://www.globalsources.com/api/gsol-skc-bff/sourcing-digest/rss