Who Gets Paid When AI Data Centers Are Built? The Companies Behind Power, Cooling, and the Grid

AI companies receive most of the attention. Much of the spending goes somewhere else.

Before an AI data center can operate, someone must secure the land, connect the grid, supply the electrical equipment, install the cables, cool the servers, and sell the electricity.

Each step creates revenue for a different group of companies.

The AI data center boom is not one investment theme. It is a chain of contracts.

This article maps that chain. It does not recommend any stock. These companies were selected to represent different parts of the AI data center value chain. They are large, publicly traded businesses with documented data center exposure, but they are not a ranking of the world’s twelve best companies.

Follow the Money, Not Only the AI Brand

A large data center project starts with capital spending.

The owner pays for buildings, grid access, transformers, cables, cooling, backup systems, and construction. After the site opens, it keeps paying for electricity, maintenance, network connections, and rent.

That creates two kinds of opportunity:

  • Project revenue from equipment and construction.
  • Recurring revenue from power, rent, connections, and service.

The difference matters. Project revenue can rise quickly and then fall. Recurring revenue is often steadier, but it may require more capital and debt.

1. Data Center Platforms: Equinix and Digital Realty

Equinix and Digital Realty own and operate data center platforms.

They sell secure space, power capacity, and connections to cloud providers, networks, and companies. The customer may install its own servers, while the operator provides the building and critical infrastructure.

Equinix is especially known for interconnection. Customers pay to connect directly with other companies, networks, and cloud services inside its ecosystem. Digital Realty combines colocation with large hyperscale campuses.

The attraction is recurring contract revenue. The risk is capital intensity. New facilities need land, power, construction, and financing before they become fully occupied.

Watch: leasing, occupancy, available power, development spending, debt, and cash flow.

2. Power and Cooling: Vertiv and Schneider Electric

Servers cannot run on a normal building power system.

Data centers need UPS systems, power distribution, switchgear, controls, batteries, and cooling. Vertiv and Schneider Electric sell many of these systems.

AI makes this layer more important. More computing can mean higher power density inside each rack. That raises the need for liquid cooling, stronger power delivery, and better controls.

Schneider Electric reported that data center strength helped lead growth in its Energy Management business in the first quarter of 2026. Vertiv has continued to expand its liquid-cooling products and services.

These companies can benefit from both new construction and later upgrades. But investors should ask whether strong orders are improving profit margins and cash flow.

Watch: orders, backlog, liquid-cooling exposure, margins, production capacity, and service revenue.

3. Grid Equipment: GE Vernova and Siemens Energy

The data center may be ready before the grid is ready.

New load can require substations, transformers, high-voltage equipment, grid controls, and sometimes new generation.

GE Vernova and Siemens Energy sell equipment used across power generation and transmission. They sit closer to the grid bottleneck than the server rack.

GE Vernova said its Electrification segment booked $2.4 billion of equipment orders to support data centers in the first quarter of 2026. That was more than its data center equipment orders for all of 2025.

Siemens Energy also reported several U.S. data-center-related orders worth a high triple-digit-million-euro amount in its first fiscal quarter of 2026.

This is an important signal. Announced campuses are becoming real equipment orders.

The risk is execution. Long backlogs can create pricing power, but they can also create delivery pressure, cost inflation, and customer delays.

Watch: book-to-bill, backlog, transformer and turbine capacity, delivery time, margins, and cancellation risk.

4. Cables and Construction: Prysmian and Quanta Services

Equipment does not install itself.

Prysmian supplies power and fiber cables. Both are important. A data center needs electricity, but it also needs fast connections to the outside world.

Quanta Services designs, builds, upgrades, and maintains energy and communications infrastructure. Its work can include transmission, substations, power generation, and large-load connections.

Prysmian has been expanding data center cable capacity and emphasizes that it can provide both digital and energy connections. Quanta reported a record $48.5 billion backlog at the end of the first quarter of 2026.

A large backlog can support future revenue. It is not the same as guaranteed profit. Labor shortages, fixed-price contracts, delays, and cost overruns can reduce returns.

Watch: backlog quality, skilled labor, contract type, project margins, raw-material costs, and free cash flow.

5. Electricity Supply: Constellation and NextEra Energy

A data center does not buy electricity once. It buys electricity every hour it operates.

This creates long-term opportunities for power producers and utilities. It also creates large capital needs.

Constellation sells electricity from a large nuclear and generation fleet. In 2026, a Constellation business signed a 380 MW agreement to serve a CyrusOne data center beside a Texas power plant. The company said its Texas agreements with CyrusOne totaled more than 1,100 MW.

NextEra Energy combines regulated utility operations with renewable generation, storage, transmission, and large-load development. Its data center strategy connects land, grid access, generation, and storage.

Rising electricity demand does not guarantee rising shareholder returns. The company must recover the cost of new assets, manage fuel and construction risk, and obtain regulatory approval where required.

Watch: contracted MW, power prices, generation cost, regulatory recovery, capital spending, debt, and project timing.

6. Backup Power and Thermal Systems: Caterpillar and Trane Technologies

Data centers need electricity even when the grid fails. They also need to remove heat every minute.

Caterpillar supplies standby and prime-power systems. It is also working on larger power solutions that combine natural gas generation, batteries, and other technologies.

Trane Technologies supplies cooling, controls, and thermal-management systems. It completed the acquisition of LiquidStack in 2026, adding direct-to-chip and immersion liquid-cooling technology.

This layer can benefit as AI racks become denser. It can also benefit from replacement and maintenance work after the original data center opens.

But both companies serve many markets. Investors must check whether data center growth is large enough to change the results of the whole company.

Watch: data center revenue exposure, equipment orders, service contracts, liquid-cooling adoption, and segment margins.

12 Companies Across the Value Chain

Layer Companies How they earn revenue Main investor risk
Data center platforms Equinix, Digital Realty Rent, power capacity, and interconnection Debt, capital spending, and slow leasing
Power and cooling Vertiv, Schneider Electric UPS, power distribution, controls, and cooling Competition and margin pressure
Grid equipment GE Vernova, Siemens Energy Generation, transformers, and high-voltage equipment Long projects and execution risk
Cables and construction Prysmian, Quanta Services Cable sales and infrastructure contracts Labor, materials, and cost overruns
Electricity supply Constellation, NextEra Energy Power sales and long-term contracts Regulation, debt, and construction cost
Backup and thermal systems Caterpillar, Trane Technologies Generators, cooling equipment, controls, and service Limited data center exposure inside a larger company

A Good Company Is Not Always a Good Stock

This is the most important investment lesson in the article.

A company may have strong orders and a valuable product. Its stock can still disappoint.

The reason may be simple:

  • the expected growth is already reflected in the price
  • the company needs too much debt or new capital
  • revenue grows but margins fall
  • backlog is delayed or canceled
  • cash flow does not follow reported profit
  • data centers are only a small part of the company

Industry growth answers one question:

Is demand increasing?

Investment analysis must answer another:

How much of that demand will reach each share?

Seven Numbers Investors Should Check

  1. Data center exposure: How much of total revenue actually comes from this market?
  2. Orders: Are customer plans becoming signed contracts?
  3. Backlog: How much future work is already booked?
  4. Book-to-bill: Are new orders arriving faster than revenue is being delivered?
  5. Margins: Is demand improving profit per dollar of sales?
  6. Free cash flow and debt: Can the company fund growth without weakening its balance sheet?
  7. Valuation: How much future success is already included in the stock price?

These numbers allow readers to compare a landlord, an equipment maker, a contractor, and a power producer without pretending that they are the same business.

The Main Idea

The most visible AI company may not receive the most reliable part of the infrastructure spending. Follow the bottleneck, the contract, and the cash flow.

The AI data center boom creates demand across many industries. Some companies sell one-time equipment. Some build projects. Some collect recurring rent or power revenue.

The best place to start is not a list of stock tickers. It is the value chain.

Once the chain is clear, investors can ask which layer has the strongest demand, the slowest supply response, the best margins, and the most reasonable price.

Related Articles

Sources

The company logos in the featured image are used only for editorial identification. Their inclusion does not imply affiliation, endorsement, or an investment recommendation.