Global Data Center Power Demand Could Reach 945 TWh: Who Benefits From the Next Wave of Electricity Investment?

 

AI Is Turning Electricity Into a Strategic Resource

Artificial intelligence has become one of the most important investment themes of the decade. Much of the attention has focused on semiconductor manufacturers, cloud providers, and software companies that are driving the rapid adoption of AI technologies. However, beneath the excitement surrounding large language models and data center expansion lies a less visible but increasingly critical factor:

Electricity.

As AI workloads become more complex, data centers require significantly more power to support training and inference activities. This growing demand is transforming electricity from a utility expense into a strategic resource. According to Gartner, global data center electricity consumption is expected to rise from approximately 415 terawatt-hours (TWh) in 2025 to around 565 TWh in 2026, with longer-term projections suggesting demand could reach approximately 945 TWh by 2030.

This trend is attracting growing attention from governments, utilities, infrastructure providers, and corporate decision-makers. The next phase of AI growth may depend not only on computing power, but also on the ability to generate, deliver, and manage the electricity required to support it.

 

Rising Power Consumption Is Creating New Infrastructure Demands

The relationship between AI and electricity demand is relatively straightforward. As AI adoption expands, companies continue investing in larger and more sophisticated data centers. These facilities require substantial amounts of electricity to power servers, cooling systems, networking equipment, and supporting infrastructure.

In the United States, the Energy Information Administration has projected electricity consumption to reach record levels in both 2026 and 2027. While multiple factors contribute to rising demand, the rapid expansion of data centers has become one of the most important structural drivers.

The implications extend far beyond utility bills. Increasing electricity demand requires new investments across the entire power ecosystem, including generation capacity, transmission infrastructure, grid modernization, and electrical equipment. As a result, AI is no longer simply a technology story. It is increasingly becoming an infrastructure story.

This shift is particularly important because power infrastructure projects often require years of planning, construction, and regulatory approval. Even if AI demand continues to grow rapidly, the supporting electricity network cannot be expanded overnight. This creates both opportunities and challenges for businesses operating throughout the energy value chain.

 

The Power Industry Is Emerging as an Unexpected AI Beneficiary

While investors often focus on companies directly involved in artificial intelligence, recent developments suggest that the power sector may become one of the most significant indirect beneficiaries of AI expansion.

Evidence of this trend is already beginning to appear across the industry. Utility mergers and acquisitions have accelerated as companies position themselves for higher electricity demand. Grid modernization initiatives are receiving increased attention, while investments in power generation and transmission infrastructure continue to expand.

The industry is also confronting growing concerns about power availability. Several studies have highlighted the possibility that electricity infrastructure could become a bottleneck for future data center growth. In some regions, developers are already facing delays related to grid connections, transformer availability, and transmission capacity.

These developments suggest that the economic impact of AI extends far beyond software and hardware. Companies that help produce electricity, deliver electricity, and support electrical infrastructure may become increasingly important participants in the broader AI ecosystem.

 

Growing Electricity Demand Is Exposing Multiple Infrastructure Bottlenecks

One of the most important insights emerging from recent industry developments is that rising electricity demand creates challenges at multiple levels of the power system.

The first challenge involves power generation. Data centers require reliable electricity around the clock, creating demand for stable sources of power capable of supporting continuous operations.

The second challenge involves transmission infrastructure. Generating electricity is only part of the equation. Electricity must also be transported efficiently from power plants to industrial users and data centers. Many existing transmission networks were not designed to accommodate the level of demand growth now being projected.

The third challenge involves critical electrical equipment. Transformers, substations, switchgear, and grid management technologies are becoming increasingly important as utilities attempt to modernize aging infrastructure and support higher electricity loads.

Because these bottlenecks exist simultaneously, investment opportunities are emerging across multiple layers of the electricity ecosystem rather than within a single industry segment.

 

Several Companies Are Positioned Across Different Layers of the AI-Electricity Value Chain

Among publicly traded companies, several businesses appear particularly well positioned to benefit from growing electricity demand associated with AI infrastructure.

Constellation Energy represents one of the most direct exposures to the power generation side of the equation. The company reported first-quarter revenue of $11.12 billion, exceeding expectations by more than 23%, while maintaining its full-year earnings guidance. Management also projected significant free cash flow generation over the coming years, reflecting confidence in the long-term demand outlook. As one of the largest providers of carbon-free electricity in the United States, Constellation occupies an important position within the power generation layer of the AI value chain.

Quanta Services represents a different type of opportunity. Rather than generating electricity, the company focuses on building and maintaining the infrastructure required to deliver it. During the first quarter of 2026, Quanta reported record backlog levels of approximately $48.5 billion alongside strong organic revenue growth. The company has also been investing in transformer manufacturing and supply chain capabilities, areas that have become increasingly important as utilities attempt to expand and modernize power networks.

GE Vernova occupies perhaps the broadest position within the electricity ecosystem. The company reported a 71% year-over-year increase in orders, while backlog expanded to approximately $163 billion. Its book-to-bill ratio reached 2.0, indicating exceptionally strong demand for its products and services. Through its involvement in power generation technologies, grid equipment, electrification systems, and transmission infrastructure, GE Vernova is positioned across multiple layers of the AI-electricity value chain.

 

The Next Wave of AI Investment May Depend on Physical Infrastructure

The first phase of the AI investment cycle was largely driven by computing power. Semiconductor companies, cloud providers, and software platforms captured most of the market's attention as businesses rushed to adopt artificial intelligence technologies.

The next phase may look very different.

As global data center electricity demand moves toward projected levels approaching 945 TWh, physical infrastructure is becoming increasingly important. Electricity generation, grid expansion, transmission networks, and electrical equipment are emerging as critical enablers of future AI growth.

For corporate leaders, investors, and strategic decision-makers, this development represents an important shift in perspective. The long-term winners of the AI era may not be limited to the companies building intelligent software. They may also include the businesses responsible for supplying the electricity and infrastructure that make large-scale AI deployment possible.

In that sense, the future of artificial intelligence may depend as much on power systems as it does on processing power.

Disclaimer :

This report is provided for informational and educational purposes only and should not be considered investment, financial, legal, or business advice. The analysis is based on publicly available information and reflects market conditions and assumptions at the time of publication, which may change without notice.