AI’s power boom is reshaping the stock market

View the chart data

Global annual data center electricity use

Year	TWh	Status
2025	485	IEA estimate
2030	950	Central projection

Every AI data center needs a connection to the physical economy: electricity to run its servers, equipment to manage that power, and cooling to keep the hardware working. That creates a spending chain that reaches well beyond the largest technology companies.

The market question is how much of that spending becomes durable earnings, and how much investors have already priced in. For SpecStocks, the useful angle is to follow the evidence into suppliers, smaller industrial companies, and measurable business developments.

There is a timely economic reason to pay attention. In a September 28 speech, Federal Reserve Governor Lisa D. Cook warned that AI investment could spread price pressure through shared inputs such as energy and construction labor. She also described the possibility of later productivity gains, with considerable uncertainty about their timing. Those were her views, rather than a promise about the path of interest rates. [1]

Why power has become a market catalyst

The IEA’s updated outlook points to electricity consumption roughly doubling between 2025 and 2030. It also identifies grid connections, equipment capacity, chips, and financing as constraints. That combination can strengthen demand for suppliers while delaying the projects those suppliers expect to serve. [2]

For a company selling electrical equipment, a capacity expansion can mean more orders and potentially stronger pricing. For a data center operator, the same expansion means upfront capital costs and the need to earn a return on them. Utilities face a different test: customer commitments, financing costs, and the regulatory treatment of new infrastructure.

Our interpretation is that the spending cycle could spread investor attention across several industries. A connection to AI is only the beginning of the analysis. The decisive questions are which company has a real role, what it can earn, and whether its stock price allows for the risks.

Follow the spending into company results

Reported results give the theme something concrete to stand on. These companies illustrate different parts of the chain. Their figures cover the stated businesses and periods; they are not a calculation of AI-only revenue.

These examples explain business exposure. They do not establish that any of the stocks is undervalued. A valuation case needs a current share price, a verified share count, debt and cash, realistic earnings or cash-flow assumptions, and a comparison with the expectations already embedded in the market.

Growth, inflation, and the cost of power

Building data centers and the infrastructure around them creates demand for construction, manufacturing, and equipment. Federal Reserve staff research notes that AI-related investment has contributed to growth, while equipment imports can offset part of the measured domestic GDP contribution. The authors also flag the difficulty of isolating AI spending from broader investment categories. [7]

“Data-center investment relies on inputs, like construction labor and energy, that are broadly used in many sectors in the economy.”

Cook’s argument connects the theme to a broad market issue: spending can arrive before the productivity benefits. Our market interpretation is that stronger equipment demand may support some suppliers’ earnings, while higher input and financing costs can pressure other businesses. The impact on household electricity bills also depends on local supply and how new infrastructure costs are allocated.

That makes interest rates relevant on both sides of the story. Borrowing costs affect the economics of infrastructure projects. They also affect the value investors place on profits expected years from now. Strong demand can coexist with a stock losing value if execution disappoints or the market lowers the multiple it is willing to pay.

Where the SpecStocks lens comes in

SpecStocks aims for undervalued and undiscovered gems with near-term in-play potential. This theme creates a research route into specialist equipment makers, cooling suppliers, electrical contractors, and related industrial businesses whose exposure may be less obvious.

Powell’s reported award shows what a useful lead looks like: a named supplier, a disclosed order, and a defined business role. Its wider portfolio also includes other industrial markets. A large award is not proof of recurring earnings, immediate revenue, or an attractive entry price. [6]

For a smaller company, the next earnings report can test whether new orders are turning into shipments, margins, and operating cash. Customer concentration, working-capital needs, debt, and potential dilution can matter as much as the size of the opportunity. A thinly traded share can also move sharply in either direction.

The editorial opportunity is to identify a business change before it becomes widely understood, then examine the valuation and the evidence that could disprove the case. Any company considered for a separate SpecStocks stock pick would need that full, current review.

What could change the market’s expectations next

• Spending plans: revisions to large customers’ infrastructure budgets can alter suppliers’ order expectations. Compare the period, cash spending, and lease treatment.
• Order conversion: a growing backlog becomes more meaningful when shipments, margins, and collections follow. Delays and cancellations can change that picture.
• Power availability: connection milestones and equipment deliveries can determine when a project begins producing revenue.
• Valuation: compare changes in a company’s earnings prospects with changes in its share price. A widely recognized theme may already command a demanding premium.
• Efficiency and adoption: lower energy use per task and uncertain usage growth can change the scale of future infrastructure demand.

AI’s power demand gives the market a set of concrete things to measure: orders, equipment capacity, completed projects, and cash generation. For SpecStocks readers, those observations are the starting point for finding overlooked company developments and judging what they may mean for shareholders.