The rapid expansion of artificial intelligence data centers in the United States is creating a growing source of demand for manufacturers far removed from the technology companies building the facilities. Generator makers, electrical equipment producers, cooling-system manufacturers, steel suppliers, cable companies and specialised industrial firms are receiving more orders as technology companies spend heavily on the physical infrastructure required to support artificial intelligence.
The effect is becoming visible in company investment, hiring and factory expansion. Generac, traditionally associated with backup generators for homes and businesses, is committing $250 million by the end of next year to equip several factories to produce larger generators for data centers. Its backlog for those products has reached about $1.6 billion, while the company expects to add roughly 1,000 workers. The development illustrates how demand from a relatively concentrated group of technology customers can move through several layers of the manufacturing economy.
Recent industry research indicates that this is not limited to generators. Electrical transformers, switchgear, cooling equipment, cables, construction materials and power-generation systems are all facing stronger demand as data center construction accelerates. The International Energy Agency estimates that capital spending by the largest technology companies exceeded $400 billion in 2025 and could rise by another 75% in 2026, although it also warns that not every planned data center project will ultimately be completed.
Power Equipment Is Emerging As A Key Constraint
Electricity is at the centre of the supply-chain effect because artificial intelligence data centers require unusually large and reliable power supplies. A facility cannot operate simply because a site has been selected and servers have been purchased. Developers also need transformers, switchgear, generators, power-conversion equipment and other infrastructure capable of delivering electricity continuously and safely.
This requirement is creating pressure on manufacturers that were already dealing with long production cycles. Large power transformers can take years to manufacture, while supplies of specialised electrical steel and other components remain constrained. Utilities and data center developers have consequently been placing orders well in advance, increasing visibility for manufacturers but also making it harder for new projects to obtain equipment quickly.
The pressure extends beyond individual factories. United States electricity demand has begun rising faster after years of relatively modest growth, with data centers among the factors contributing to the increase. Federal energy projections indicate that electricity consumption by data center servers could reach between 446 billion and 818 billion kilowatt-hours annually by 2050, depending on how quickly computing demand and infrastructure expand.
That creates a chain of industrial demand. More computing capacity requires more electricity, greater electricity demand requires additional generation and grid equipment, and that equipment requires more metals, electrical components, machinery and manufacturing capacity. The economic impact therefore reaches companies that may never sell anything directly to an artificial intelligence developer.
Manufacturers Are Following The Orders
The effect is particularly visible among industrial companies that have historically supplied several unrelated sectors. Timken, for example, supplies engineered bearings to industries including aerospace and defence, but its management has identified data center construction as another source of demand because the facilities require large buildings, power systems, roads and other infrastructure.
Other manufacturers are responding more directly. Siemens has announced more than $200 million of investment in new United States plants producing electrical equipment for data centers and industrial customers. The strategy reflects the opportunity created by the current investment cycle while also limiting exposure by maintaining business outside the data center market.
Smaller manufacturers are seeing similar effects. Southeastern Hose, which traditionally supplied steel and petrochemical customers, has reported a sharp increase in data center-related demand and significant revenue growth. Such companies demonstrate how the effect can travel through the supply chain even when the original manufacturer is several stages removed from the technology industry.
The broader manufacturing data provide some evidence of stronger activity, although they do not establish that artificial intelligence is responsible for the entire improvement. United States manufacturing employment increased during the first seven months of 2026 after substantial job losses in 2025, while factory activity reached its highest level in several years. Semiconductor plant construction and other large industrial projects are also contributing to the improvement.
That distinction is important. The current industrial recovery is not uniform, and many manufacturers remain exposed to weaker consumer and housing demand. Generac itself provides an example: its traditional household generator business remains under pressure while data center demand is exceptionally strong. The result is a manufacturing sector in which some companies are experiencing rapid expansion while others continue to face difficult conditions.
Shortages Are Raising The Cost Of Expansion
Strong orders are creating an opportunity for manufacturers, but they are also exposing limits in production capacity. Industry estimates indicate that the United States data center electrical equipment market could increase sharply through 2030, driven by demand for transformers, switchgear and power distribution systems. Manufacturers must therefore decide how much additional capacity to build without knowing how long the current pace of data center investment will last.
The supply problem is particularly difficult because many of the required products cannot be produced quickly. Large transformers, turbines and specialised electrical equipment involve complex manufacturing processes, specialised materials and skilled labour. Increasing capacity can therefore take years, even when companies have sufficient capital and strong customer demand.
That has already affected prices and delivery schedules. Some equipment manufacturers have been revisiting older orders because the cost of producing and delivering equipment has increased since those contracts were signed. Customers building data centers can be willing to accept higher prices because delays in obtaining power equipment can postpone the opening of facilities that require enormous amounts of capital.
The result is a market in which manufacturers have greater pricing power but also face pressure to invest quickly. If they expand too slowly, they may lose customers and allow competitors to gain market share. If they expand too aggressively and data center investment slows, they could be left with factories and equipment designed around demand that no longer exists.
Companies Are Trying To Limit Bubble Risk
The possibility of a slowdown is becoming an important consideration for manufacturers because artificial intelligence investment has reached a scale that makes the sector increasingly important to their revenues. Companies that previously served a wide range of industries are now devoting more production capacity to data center customers, creating opportunities but also greater exposure to changes in technology spending.
Some manufacturers are reducing that risk through long-term contracts. Siemens, for example, has used multiyear agreements that include financial penalties if customers fail to meet commitments. Such arrangements provide manufacturers with greater confidence before making large investments and ensure that some of the risk is shared with the companies demanding additional capacity.
Other manufacturers are maintaining a broader customer base. That approach may be particularly important for smaller suppliers that cannot easily absorb a sharp decline in orders. Businesses with established customers in steel, petrochemicals, construction or other industries have an incentive to preserve those relationships even while accepting new data center work.
The concern is not that a slowdown is inevitable. Current demand remains strong, and major technology companies continue to commit substantial capital to computing infrastructure. The uncertainty lies in how long that spending can continue at its current pace and whether improvements in computing efficiency, changes in artificial intelligence models or weaker investment returns could alter future infrastructure requirements.
The Boom Is Extending Into Industrial Real Estate
The supply-chain effect is also becoming visible in industrial property markets. Companies manufacturing generators, transformers, cooling systems and other data center equipment need warehouses and production facilities, often close to skilled workers and transportation networks rather than beside the data centers themselves.
Research covering major United States data center markets found that businesses connected to the sector accounted for 10.4% of new industrial leasing activity between 2022 and 2025. That share increased to 14.4% in 2025 as artificial intelligence investment accelerated. The figures indicate that the economic activity associated with data centers extends well beyond the facilities themselves.
This broader demand could become an important part of the industrial investment cycle, particularly in regions where electricity availability, transport infrastructure and manufacturing capacity can support both data center development and its supplier network. But it also creates pressure on existing infrastructure, including electricity grids, industrial land and skilled labour.
The emerging pattern is therefore less about artificial intelligence replacing traditional manufacturing than about technology investment becoming an additional customer for parts of the industrial economy. Generators, transformers, cooling systems, cables, bearings, hoses and construction materials remain conventional industrial products, but demand from data centers is increasing their importance.
Whether that demand produces a sustained manufacturing expansion will depend on the durability of artificial intelligence investment. For now, the evidence shows that the data center build-out is generating orders well beyond servers and chips, while manufacturers are responding with new factories, additional workers and greater production capacity. The central challenge is determining how much capacity the market can absorb without turning today's exceptional demand into tomorrow's excess supply.
(Source:www.reuters.com)
The effect is becoming visible in company investment, hiring and factory expansion. Generac, traditionally associated with backup generators for homes and businesses, is committing $250 million by the end of next year to equip several factories to produce larger generators for data centers. Its backlog for those products has reached about $1.6 billion, while the company expects to add roughly 1,000 workers. The development illustrates how demand from a relatively concentrated group of technology customers can move through several layers of the manufacturing economy.
Recent industry research indicates that this is not limited to generators. Electrical transformers, switchgear, cooling equipment, cables, construction materials and power-generation systems are all facing stronger demand as data center construction accelerates. The International Energy Agency estimates that capital spending by the largest technology companies exceeded $400 billion in 2025 and could rise by another 75% in 2026, although it also warns that not every planned data center project will ultimately be completed.
Power Equipment Is Emerging As A Key Constraint
Electricity is at the centre of the supply-chain effect because artificial intelligence data centers require unusually large and reliable power supplies. A facility cannot operate simply because a site has been selected and servers have been purchased. Developers also need transformers, switchgear, generators, power-conversion equipment and other infrastructure capable of delivering electricity continuously and safely.
This requirement is creating pressure on manufacturers that were already dealing with long production cycles. Large power transformers can take years to manufacture, while supplies of specialised electrical steel and other components remain constrained. Utilities and data center developers have consequently been placing orders well in advance, increasing visibility for manufacturers but also making it harder for new projects to obtain equipment quickly.
The pressure extends beyond individual factories. United States electricity demand has begun rising faster after years of relatively modest growth, with data centers among the factors contributing to the increase. Federal energy projections indicate that electricity consumption by data center servers could reach between 446 billion and 818 billion kilowatt-hours annually by 2050, depending on how quickly computing demand and infrastructure expand.
That creates a chain of industrial demand. More computing capacity requires more electricity, greater electricity demand requires additional generation and grid equipment, and that equipment requires more metals, electrical components, machinery and manufacturing capacity. The economic impact therefore reaches companies that may never sell anything directly to an artificial intelligence developer.
Manufacturers Are Following The Orders
The effect is particularly visible among industrial companies that have historically supplied several unrelated sectors. Timken, for example, supplies engineered bearings to industries including aerospace and defence, but its management has identified data center construction as another source of demand because the facilities require large buildings, power systems, roads and other infrastructure.
Other manufacturers are responding more directly. Siemens has announced more than $200 million of investment in new United States plants producing electrical equipment for data centers and industrial customers. The strategy reflects the opportunity created by the current investment cycle while also limiting exposure by maintaining business outside the data center market.
Smaller manufacturers are seeing similar effects. Southeastern Hose, which traditionally supplied steel and petrochemical customers, has reported a sharp increase in data center-related demand and significant revenue growth. Such companies demonstrate how the effect can travel through the supply chain even when the original manufacturer is several stages removed from the technology industry.
The broader manufacturing data provide some evidence of stronger activity, although they do not establish that artificial intelligence is responsible for the entire improvement. United States manufacturing employment increased during the first seven months of 2026 after substantial job losses in 2025, while factory activity reached its highest level in several years. Semiconductor plant construction and other large industrial projects are also contributing to the improvement.
That distinction is important. The current industrial recovery is not uniform, and many manufacturers remain exposed to weaker consumer and housing demand. Generac itself provides an example: its traditional household generator business remains under pressure while data center demand is exceptionally strong. The result is a manufacturing sector in which some companies are experiencing rapid expansion while others continue to face difficult conditions.
Shortages Are Raising The Cost Of Expansion
Strong orders are creating an opportunity for manufacturers, but they are also exposing limits in production capacity. Industry estimates indicate that the United States data center electrical equipment market could increase sharply through 2030, driven by demand for transformers, switchgear and power distribution systems. Manufacturers must therefore decide how much additional capacity to build without knowing how long the current pace of data center investment will last.
The supply problem is particularly difficult because many of the required products cannot be produced quickly. Large transformers, turbines and specialised electrical equipment involve complex manufacturing processes, specialised materials and skilled labour. Increasing capacity can therefore take years, even when companies have sufficient capital and strong customer demand.
That has already affected prices and delivery schedules. Some equipment manufacturers have been revisiting older orders because the cost of producing and delivering equipment has increased since those contracts were signed. Customers building data centers can be willing to accept higher prices because delays in obtaining power equipment can postpone the opening of facilities that require enormous amounts of capital.
The result is a market in which manufacturers have greater pricing power but also face pressure to invest quickly. If they expand too slowly, they may lose customers and allow competitors to gain market share. If they expand too aggressively and data center investment slows, they could be left with factories and equipment designed around demand that no longer exists.
Companies Are Trying To Limit Bubble Risk
The possibility of a slowdown is becoming an important consideration for manufacturers because artificial intelligence investment has reached a scale that makes the sector increasingly important to their revenues. Companies that previously served a wide range of industries are now devoting more production capacity to data center customers, creating opportunities but also greater exposure to changes in technology spending.
Some manufacturers are reducing that risk through long-term contracts. Siemens, for example, has used multiyear agreements that include financial penalties if customers fail to meet commitments. Such arrangements provide manufacturers with greater confidence before making large investments and ensure that some of the risk is shared with the companies demanding additional capacity.
Other manufacturers are maintaining a broader customer base. That approach may be particularly important for smaller suppliers that cannot easily absorb a sharp decline in orders. Businesses with established customers in steel, petrochemicals, construction or other industries have an incentive to preserve those relationships even while accepting new data center work.
The concern is not that a slowdown is inevitable. Current demand remains strong, and major technology companies continue to commit substantial capital to computing infrastructure. The uncertainty lies in how long that spending can continue at its current pace and whether improvements in computing efficiency, changes in artificial intelligence models or weaker investment returns could alter future infrastructure requirements.
The Boom Is Extending Into Industrial Real Estate
The supply-chain effect is also becoming visible in industrial property markets. Companies manufacturing generators, transformers, cooling systems and other data center equipment need warehouses and production facilities, often close to skilled workers and transportation networks rather than beside the data centers themselves.
Research covering major United States data center markets found that businesses connected to the sector accounted for 10.4% of new industrial leasing activity between 2022 and 2025. That share increased to 14.4% in 2025 as artificial intelligence investment accelerated. The figures indicate that the economic activity associated with data centers extends well beyond the facilities themselves.
This broader demand could become an important part of the industrial investment cycle, particularly in regions where electricity availability, transport infrastructure and manufacturing capacity can support both data center development and its supplier network. But it also creates pressure on existing infrastructure, including electricity grids, industrial land and skilled labour.
The emerging pattern is therefore less about artificial intelligence replacing traditional manufacturing than about technology investment becoming an additional customer for parts of the industrial economy. Generators, transformers, cooling systems, cables, bearings, hoses and construction materials remain conventional industrial products, but demand from data centers is increasing their importance.
Whether that demand produces a sustained manufacturing expansion will depend on the durability of artificial intelligence investment. For now, the evidence shows that the data center build-out is generating orders well beyond servers and chips, while manufacturers are responding with new factories, additional workers and greater production capacity. The central challenge is determining how much capacity the market can absorb without turning today's exceptional demand into tomorrow's excess supply.
(Source:www.reuters.com)