Data Center Power Demand Poised for Steep Growth, and Two Industrial Suppliers Are in Focus

Artificial intelligence is reshaping electricity consumption patterns in the United States, and the numbers involved are substantial. Research from Goldman Sachs released Sept. 1, 2026, projects that U.S. data center power demand could climb to roughly 108 gigawatts by 2030 — a figure that underscores how quickly AI workloads are straining the nation’s electrical infrastructure.

The projected surge has implications across the power value chain. Utilities responsible for generation and grid delivery face growing load growth after years of flat demand, while equipment providers operating inside the data center walls are seeing increased attention for their role in power management and thermal control.

Inside-the-Data-Center Players

One company positioned in the middle of this buildout is Vertiv Holdings Co (NYSE: VRT), an industrial firm that designs, manufactures, and services critical digital infrastructure for data centers, communication networks, and commercial and industrial environments across the Americas, Asia Pacific, Europe, the Middle East, and Africa. Its portfolio addresses exactly the two challenges AI-driven facilities face most acutely: delivering reliable power to racks and removing the heat those racks generate.

Market activity reflects the sector’s momentum. Vertiv shares traded at $280.53, up 4.32% from the previous close of $268.91, giving the company a market capitalization of roughly $87.6 billion. The stock sits in the industrials sector, within the electrical equipment and parts industry.

The Utility Angle

On the grid side, American Electric Power (NASDAQ: AEP) represents the utility category that analysts have highlighted as a beneficiary of data center expansion. As hyperscale operators sign large power contracts, regulated utilities in data-center-heavy regions are seeing load forecasts revised upward — a dynamic the Goldman Sachs research helps quantify.

Why the Power Theme Matters Now

The AI power buildout has become one of the more closely followed stories in industrials and utilities alike. Training and inference workloads for large AI models consume significantly more energy per facility than traditional computing deployments, and hyperscalers have responded with aggressive capacity announcements. That, in turn, drives orders for switchgear, power distribution units, uninterruptible power supplies, and liquid cooling systems — categories where Vertiv competes — while utilities plan new generation and transmission capacity to serve the demand.

The Goldman Sachs projection of near-108 gigawatts of U.S. data center demand by 2030 provides a common reference point for how large the opportunity set could be, though actual buildout will depend on grid interconnection timelines, permitting, and the pace of AI infrastructure spending.

Source: original release

What to watch

• Vertiv’s upcoming quarterly earnings report, including order flow and backlog commentary tied to data center demand.
• Utility capital expenditure plans and load-growth disclosures, including updates from American Electric Power on data center contracts.
• Revisions to third-party forecasts of U.S. data center power demand as AI capacity announcements evolve.
• Grid interconnection and transmission project timelines that could affect how quickly new data center capacity comes online.