The explosion of AI is not limited to enriching giants in semiconductors, software, and digital platforms. Instead, it is uplifting the fortunes of hundreds of industrial, utility, and mining companies that until recently appeared as mature, traditional, and unattractive to investors.
As the Financial Times, which dedicates one of its reports to the phenomenon, writes, these companies are providing the physical infrastructure to build, power, and cool the massive data centers necessary for AI development and operation.
Investors, seeking concrete ways to ride the AI megatrend, are increasingly looking beyond chips and algorithms, focusing on companies that make the underlying physical construction possible.
The result is that many of these listed companies are posting stock market performances superior to the global index, in a phenomenon injecting new vitality into sectors long considered static or in decline.
An impressive market performance
According to the analysis conducted by the Financial Times, more than 200 public companies linked to data center or semiconductor supply chains have outperformed the MSCI World index, which itself closed the year on June 9 with a solid increase of over 21%.
Among the most significant examples are Caterpillar, known primarily for construction machinery but now a supplier of generators for data centers; the German Hochtief, a historic engineering company with 150 years of activity soon to enter the Dax index; and Nucor, a steel producer that spoke of a “red-hot demand” capable of generating a “real tsunami of profits.”
Ford also saw its shares rise by a fifth in May after announcing a shift of its business from electric vehicles toward battery storage dedicated to data centers.
Infrastructure demand
Industry analysts describe the last six months as a period of explosive growth.
Alex Cordovil of Dell’Oro calls it a “very exciting” time for the industry, while George Featherstone of Barclays highlights how many companies are registering triple-digit order growth compared to the previous year.
Centuries-old companies, once seemingly destined for slow decline, are finding new dynamism by quickly adapting to AI’s needs.
Driving this boom are massive investments from major tech operators: Alphabet, Microsoft, Amazon, Meta, and Oracle plan to spend a combined $700 billion in capital expenditures alone in 2026.
In the United States, monthly spending on data center construction hit $50 billion in April, and JLL forecasts that their global capacity will double, reaching 200 GW by 2030.
The technical complexity of data centers
Data centers require much more advanced engineering solutions than those used for simple cloud storage.
Servers must be connected more tightly and densely, greatly increasing the need for advanced cabling and optical components.
A notable case is Corning, an American company founded 175 years ago and inventor of the famous Pyrex glass, which today also supplies iPhone screens: its shares have jumped over 270 percent in the past year thanks to major contracts with Meta and Nvidia for supplying optical fibers destined for AI data centers.
Energy voracity
Training AI models consumes enormous amounts of electricity, creating unprecedented demand for power management systems, high-voltage electronics, backup solutions, and generation technologies.
American company Eaton saw data center orders increase by 240 percent in the first quarter, with its CEO Paulo Ruiz forecasting a tripling of AI energy demand over the next five years.
French company Legrand has doubled its revenues this decade, with half of the growth coming from data centers, which now represent more than a quarter of total turnover.
Even gas turbines, considered almost obsolete until a few years ago, have waiting lists up to seven years in the United States. Siemens Energy recorded its strongest quarter ever for orders in the first quarter of 2026.
Several companies—from Caterpillar to Howmet Aerospace, from Finnish Wärtsilä to Baker Hughes—have redirected part of their production toward data centers.
As Cordovil summarizes: “Anything capable of turning a motor will sooner or later end up in a data center.”
Cooling
AI processors generate intense heat, making both traditional air conditioning systems and especially liquid cooling technologies that use water to prevent chip overheating indispensable.
Shares of Comfort Systems USA, a manufacturer of air conditioning systems, have soared 260 percent in the past year. Schneider Electric acquired an $850 million stake in the specialized company Motivair.
The prominence of electric companies
Utilities are mobilizing rapidly to meet the demands of large tech groups.
In Europe, Spanish company Iberdrola is among the main providers of energy contracts for tech companies, while in the United States Entergy hit record highs after signing a $10 billion deal with Meta.
The need for constant or off-grid power is reshaping the landscape of traditional energy infrastructure.
Long-term prospects
However, there are signs of caution. According to Bain & Company, the tech industry will need to generate about $2 trillion in annual AI revenues to justify current investment levels.
Many industrial companies suffered stock declines after the launch of the Chinese model DeepSeek in early 2025, raising fears of a possible reduction in investments.
The OECD has also warned of possible “repricing” if the conflict between the United States and Iran prolongs, driving up energy costs and hindering projects in the Middle East.
Despite these concerns, most operators remain optimistic and consider this a structural, long-term investment cycle. Schneider Electric calls AI data centers a “central growth engine,” while Siemens speaks of a long-term trend, with its dedicated division growing 40 percent in 2025.
One of the most appreciated advantages is the speed at which projects proceed compared to traditional industrial clients: many AI sector customers are willing to pay premiums to secure supplies.
