In late September 2026, French group Schneider Electric unveiled a software-defined medium-voltage switchgear system designed to shorten deployment times and accelerate the expansion of AI data centers. The company plans to continue pilot programs through 2027 before scaling up supply from 2028.

Schneider's move comes as groups such as Siemens, Eaton, ABB and LS Electric are also developing products or expanding production capacity to serve AI data centres.

These developments reflect a trend receiving less attention than the semiconductor race: AI is reshaping demand for electrical equipment, from transformers and switchgear to power distribution and management systems.

Output constraints

According to the International Energy Agency (IEA), global data center electricity consumption is forecast to rise from 485 TWh in 2025 to around 950 TWh by 2030, nearly doubling in five years.

Electricity consumption by AI-focused data centers is expected to triple. The IEA also recorded a roughly 50% increase in electricity demand from this segment in 2025 alone, significantly above the 17% growth recorded across the data center sector as a whole.

But the key issue is not just total electricity consumption; it is also how servers use power.

According to Schneider Electric, data center racks traditionally consumed around 5-15 kW per rack. By contrast, configurations using next-generation GPUs for AI workloads can reach 130-140 kW per rack.

When multiple GPUs operate within a single server rack, systems must not only supply more electricity but also maintain power quality, manage load fluctuations and ensure backup capacity in the event of a failure.

At the same time, the heat generated by high-power GPUs makes conventional air-cooling methods difficult to use in some high-density configurations.

AI is therefore creating two sets of demands for the electrical equipment industry: more equipment to support the construction of new data centers, and technological upgrades to accommodate server architectures with higher power densities.

The trend is evident in investment by international groups.

In April 2026, Eaton announced an investment of more than $30 million to build a medium-voltage switchgear manufacturing facility in Nebraska, the United States. The plant is expected to begin operations in 2027. The company identified the AI data center boom as one of the drivers of demand for these products.

Siemens is also expanding its electrical equipment supply capacity. Following an investment of more than $165 million in U.S. factories announced in March 2026, the German group announced plans in July to invest 300 million euros in Germany to increase output of electrical equipment serving the energy transition and AI data centers.

On the technology front, ABB is developing new power distribution solutions for data centers, while Schneider is pursuing software-defined switchgear architectures. These developments aim to meet requirements for capacity, reliability and scalability in power systems.

Alex Cordovil, research director at Dell'Oro Group, has noted that equipment delivery times, construction labor availability, grid connections and community acceptance remain constraints on data center development.

This makes electrical equipment increasingly important to the AI infrastructure supply chain, as projects require the coordination of multiple factors, including new solutions and expanded supply capacity to meet installation schedules.

Where do Vietnamese companies stand in the new market?

In Vietnam, investment in data center infrastructure is also expanding.

According to the Ministry of Science and Technology, Vietnam had 42 commercial data centers with total designed capacity of around 372.75 MW as of August 2026, more than four times the level in 2021. This figure refers to designed capacity, not actual operating capacity or electricity consumption.

Alongside existing infrastructure, several large-scale projects are under development or in preparation.

In August 2025, Viettel broke ground on the An Khanh Data Center in Hanoi, with a total announced investment of VND17.5 trillion ($675.27 million). In Ho Chi Minh City, CMC was approved as the investor of a hyperscale data center project worth around VND6.26 trillion ($241.55 million) in September 2026.

Earlier, in April 2026, Ho Chi Minh City granted investment certificates for two data center projects, Starmason and Evolution DC VN HCMC, with registered capital of more than $480 million and $508 million, respectively.

These figures reflect the planned investment scale of projects at different stages of development and do not mean that all the capital has already been disbursed.

Nevertheless, expanding data infrastructure creates demand for power supply, distribution systems, cooling equipment and backup power sources. These are areas in which domestic electromechanical contractors and electrical equipment manufacturers could participate.

Refrigeration Electrical Engineering Corporation (REE), listed on the Ho Chi Minh Stock Exchange as REE, is one company that has already secured specific contracts linked to such projects.

According to its 2025 annual report, its subsidiary REE M&E JSC won several major projects, including the Viettel Hoang Hoa Tham Data Center.

In early 2026, REE M&E was selected to carry out another package under Viettel's An Khanh central data center project. The contract covered the supply of materials and equipment, as well as construction and installation of the electromechanical systems for the first phase of the server rooms, with a value of VND786.41 billion ($30.34 million), around 25.4% below the estimated price of nearly VND1,054 billion ($40.65 million).

The contract shows that REE has moved beyond merely exploring opportunities and has secured work on a sizeable data infrastructure project.

According to a report by Vietcap Securities following REE's extraordinary shareholders' meeting on July 10, 2026, the value of newly signed M&E contracts reached around VND5.9 trillion ($227.66 million) in the first half of the year, equivalent to 98% of its full-year 2026 target of VND6 trillion.

Vietcap estimated the division's outstanding contract backlog at around VND10 trillion ($385.87 million) at the end of June.

However, rapid growth in contract value does not necessarily translate into a corresponding increase in profit contributions.

At the extraordinary shareholders' meeting in July, REE management said revenue and profit in its M&E division in the first half had yet to reach 50% of their full-year targets. The estimated profit breakdown presented at the meeting showed that the power segment, including hydropower, thermal power and renewable energy, contributed around 56% of profit, while office leasing accounted for about 27%. The water and building M&E divisions together contributed less than 10%.

These figures were estimates presented at the time of the meeting. Nevertheless, they show that REE's earnings remain driven mainly by its traditional businesses, while building M&E has yet to play a role commensurate with the scale of its new contracts. This also indicates that AI infrastructure remains an emerging growth avenue rather than a significant earnings driver for REE.

While REE is entering the market as an M&E contractor, GELEX Electric JSC, listed on the Ho Chi Minh Stock Exchange as GEE, is approaching it through manufacturing.

GELEX Electric owns several brands across the electrical equipment supply chain, including CADIVI, which makes power cables; THIBIDI and MEEPOWER, which manufacture transformers; and EMIC, which supplies electrical metering equipment.

These product categories are linked to power transmission and distribution and are also used in electricity supply systems for data centers.

At its annual shareholders' meeting on April 8, 2026, GELEX Electric management announced a product development strategy for 2026-2030 that includes solutions for data center infrastructure, protective switchgear, smart monitoring products and smart grids.

The strategy reflects the company's view that data infrastructure is a potential growth market that can build on its existing electrical equipment manufacturing base.

However, unlike REE, publicly available information on GEE confirms its product development direction but does not yet provide clear evidence of equipment supply contracts for AI data centers or actual revenue from these product categories.

The difference between the two companies reflects two distinct stages of market participation: REE has already secured specific M&E contracts, while GEE is preparing its product portfolio to address emerging demand.

According to a forecast published by Dell'Oro Group in August 2026, global revenue for data center physical infrastructure manufacturers could reach $120 billion by 2030, representing a compound annual growth rate (CAGR) of 22% over 2025-2030. This presents a significant opportunity for manufacturers such as GEE.

However, these projects impose stringent requirements for reliability, redundancy, technical standards and compatibility between systems. Domestic suppliers must also compete with international groups that already offer specialized products and have extensive deployment experience across multiple markets.

Moreover, not every announced data center project will necessarily proceed as planned. Power availability, construction schedules, financing and demand for leased capacity can all affect when orders materialize.

The AI wave is creating new demand for the electrical equipment industry, both in terms of supply volumes and technological requirements. However, translating market demand into corporate revenue depends on manufacturing capabilities, technical standards, the ability to win contracts and project execution schedules. These factors will ultimately determine whether AI-driven opportunities translate into business results.