Artificial intelligence runs on electricity. Every AI model is trained and used in data centers packed with servers that draw power day and night and need even more power to keep cool. As AI spreads, forecasts of electricity demand from data centers have risen sharply. This has brought power, cable and cooling companies into AI conversations that used to be only about technology. This article explains why, which kinds of companies are involved, and what investors should look at.
Why AI needs so much electricity
Powerful chips use more power. AI chips do enormous numbers of calculations at once. A rack of AI servers can draw several times the electricity of a rack of ordinary servers.
Heat must be removed. Almost all the electricity used by servers turns into heat. Cooling systems, from large air-conditioning units to liquid cooling that runs close to the chips, use a significant extra share of power.
Data centers never sleep. Unlike offices or factories that slow down at night, data centers run around the clock, so they need a steady, reliable supply with backup for any outage.
The scale is growing. Technology companies are planning data-center campuses that need as much electricity as a small city. Connecting them to the grid, and finding enough generation to supply them, has become one of the main limits on how fast AI capacity can grow.
Which kinds of companies are involved
1. Power generators. Companies that produce electricity, from coal, gas, hydro, nuclear, solar or wind, can see new demand from data centers. Many data-center owners want clean energy, which supports demand for renewable power and long-term supply contracts.
2. Transmission and distribution. Electricity must travel from power plants to data centers through high-voltage lines, substations and local networks. Building and upgrading this grid is a large, long-term investment.
3. Electrical equipment makers. Transformers, switchgear, circuit breakers and backup power systems are needed at every step, from the grid to inside each data center. Demand for some of this equipment has outpaced supply in many countries.
4. Cables and wires. Power cables carry electricity into and through data centers, and fibre-optic cables carry data. Cable makers can benefit from both.
5. Cooling specialists. Companies making industrial air-conditioning, chillers and liquid-cooling systems are part of the chain, especially as AI servers produce more heat in less space.
6. Energy storage and backup. Batteries and generators keep data centers running during outages and help balance renewable power.
The Indian angle
India's electricity demand is growing for many reasons at once: economic growth, more air-conditioning, electric vehicles, manufacturing and now data centers. Data centers are still a small share of India's total power use, but they are concentrated in a few hubs such as Mumbai, Chennai and Hyderabad, where they can strain local supply. India is also adding large amounts of solar and wind capacity, and many data-center projects sign agreements to buy renewable power. This links India's power, renewable, equipment and cable companies to the AI theme, alongside their many other sources of demand.
Example on PositoAI: Tata Power
Tata Power, listed on the NSE and BSE, is one of India's largest integrated power companies. It generates electricity from thermal, hydro, solar and wind plants, distributes power to customers in several regions, and has businesses in solar manufacturing, rooftop solar and electric-vehicle charging. Data centers are one of many sources of demand for a company like this, which shows how the AI theme reaches well beyond technology firms.
On PositoAI (18 Oct 2026): Positivity Score 20,927, ranked #1,443 of 2,084 companies in India and #2,861 worldwide; market cap about ₹1.1 lakh crore; P/E 29.3; P/B 2.86. These figures change daily.
What to watch as an investor
How much demand really comes from data centers. For most power and equipment companies, data centers are one customer group among many. Company presentations sometimes show order books or contracts linked to data centers; check how large that share is.
Order books. For equipment and cable makers, the value of orders received but not yet delivered shows future revenue. Growing order books suggest strong demand; watch for delays in execution.
Capital spending and debt. Power plants and grids are expensive and take years to build. Rising debt can be manageable if projects earn steady returns, but heavy borrowing adds risk.
Regulation and tariffs. Electricity prices and returns for many power companies are set or influenced by regulators. Policy changes can affect profits more than demand does.
Fuel and input costs. Coal and gas prices for generators, and copper and aluminium prices for cable and equipment makers, affect margins.
Valuation. When a sector becomes part of a popular theme, its shares can rise ahead of profits. Compare price-to-earnings (P/E) and price-to-book (P/B) ratios with the company's own history and with peers.
Common misconceptions
"Power companies are now AI companies." Data centers add demand, but power companies' results still depend mostly on their wider customer base, fuel costs and regulation.
"More AI always means proportionally more electricity." More efficient chips, cooling and software can reduce the power needed for each task, even as total usage grows.
"Demand growth guarantees profit growth." Costs, regulated prices, execution delays and debt can all limit how much extra demand turns into profit.
Frequently asked questions
How much electricity do data centers use?
Globally, data centers use a small but fast-growing share of electricity, and in some regions with many data centers the share is much higher.
Why do data-center owners want renewable power?
Many large technology companies have clean-energy targets, and long-term renewable contracts can also give predictable power costs.
What is an order book?
The total value of orders a company has received but not yet completed, which indicates future revenue.
Why is grid connection a bottleneck?
New high-voltage lines and substations take years to plan, approve and build, so data centers can be ready before the power to run them is.
Are cable companies part of the AI theme?
Yes, indirectly: power and data cables are needed for data centers, grids and renewable projects.
How to use this on PositoAI
Search power, cable, electrical equipment and cooling companies on PositoAI to see each one's Positivity Score, rank, P/E, debt and any red or yellow flags. Use Go Deep to follow revenue, profit and debt over time, and Compare to put two companies side by side. PosiStep 13 of PosiCourse, PositoAI's free stock market course, explains cash flow, which matters a lot for capital-heavy businesses.
Suggested reading
Cooling and Power for AI: Why They Matter · Data Centers and the Stock Market · What Is a Data Center? · What Are AI Stocks? · AI Chips Explained · What Is Inflation?
Summary
AI needs large amounts of electricity because powerful chips draw more power, the heat they produce must be removed, and data centers run around the clock at growing scale. This links power generators, transmission and distribution companies, electrical equipment makers, cable makers, cooling specialists and energy-storage firms to the AI theme. For most of them, data centers are one source of demand among many, so order books, capital spending, debt, regulation, input costs and valuation matter more than the AI label.