Global Equity Pulse: Beyond the Chips—Powering the AI Revolution

Opportunities in the AI value chain

Artificial intelligence (AI) may run on chips, but it also runs on electricity. As AI develops, access to reliable power could constrain data-center growth. At TGI, we believe this broadens the opportunity beyond semiconductors to the infrastructure needed to generate, transmit and manage electricity. Less visible parts of the AI value chain, including power generation and electrical equipment, could benefit strongly from the boom.

AI Has a Growing Appetite for Electricity

The numbers illustrate the challenge. The International Energy Agency estimates that global data centers consumed around 415 terawatt-hours (TWh) of electricity in 2024, roughly 1.5% of global electricity consumption. It expects that figure to more than double to around 945 TWh by 2030, with AI the most important driver. In the United States, data centers could account for almost half of electricity-demand growth through 2030.

Other forecasts point to an even larger impact. The Electric Power Research Institute estimates that data centers could consume 9% to 17% of US electricity by 2030. That compares with roughly 4% to 5% today. Its medium-growth scenario implies annual US data center electricity consumption could rise from approximately 184 TWh in 2024 to 596 TWh in 2030.

Outlook for Electricity Demand from US Data Centers

Source: US Energy Information Administration. Data as of April 8, 2026. There is no assurance that any estimate, forecast or projection will be realized.

A Localized Challenge

The demand for power from AI is enormous, but it is also highly concentrated. Hyperscalers are developing large computing campuses that require enormous amounts of electricity. Our research notes that some data centers use more power than some small states and may challenge grids not designed to accommodate rapid increases in local demand.

That creates investment needs beyond the data center itself. New generating capacity is required. Electricity must be delivered through transmission and distribution networks. And increasingly power-intensive servers require more sophisticated electrical and cooling equipment. Where the grid cannot expand quickly enough, data center developers might also seek dedicated or behind-the-meter generation.¹

This impact is already visible in company fundamentals. TGI research points to strong demand for gas turbines, high-voltage equipment and other infrastructure linked to the data-center buildout.

Meeting the growth in electricity demand is likely to require a mix of sources. Renewables can play a significant role, while natural gas can provide reliable generation. Nuclear power may become increasingly important over time.

We believe this creates a potentially attractive new group of AI “picks and shovels." Opportunities extend from utilities and power generation to turbines, transformers, high-voltage equipment, power management and cooling.

A Global Investment Opportunity

From our bottom-up perspective, we see several ways to gain exposure to these structural trends:

• Power generation: We see an attractive long-term opportunity in a leading US power provider positioned to benefit from rising electricity demand from data centers and other large commercial customers. The investment case extends beyond AI. Years of grid modernization and additional generating capacity could support sustained capital investment and earnings growth as electricity demand rises.
• Electrical equipment and cooling: We also see opportunity in a Japanese company supplying infrastructure for next-generation data centers. Its capabilities span power systems, semiconductors and advanced cooling, positioning it to potentially benefit as AI drives higher power density and puts a greater premium on energy efficiency. Internal research highlights increasingly power-intensive AI server architectures and the need for more efficient power-delivery systems.
• Power management: Elsewhere, TGI research is identifying opportunities among more specialized suppliers. One Asian power-management company, for example, reported 70% year-on-year growth in its AI-related businesses in the first quarter of 2026 and expects AI-related products to represent approximately 30% of its 2026 revenue mix.

Strong Tailwinds, but Selectivity Remains Key

A favorable industry backdrop does not necessarily make every power-related company an attractive investment. TGI’s approach is grounded in bottom-up stock selection and price discipline, particularly when enthusiasm for a structural theme is already reflected in valuations.

The scale of AI investment is enormous, but ultimately that capital must generate acceptable returns. We therefore remain focused on companies where structural demand is supported by strong fundamentals, sustainable earnings potential and attractive long-term risk/reward.

 

Market Review Global equities were broadly unchanged in October, but this masked wide differences across regions and investment styles. Equities in the United States and Europe rose modestly, while equities in Asia Pacific and emerging markets overall declined. Equities in Japan also finished lower, although they held up better than in several technology-heavy Asian markets. Value stocks clearly outperformed growth as investors reduced exposure to crowded AI, semiconductor and momentum positions. In the United States, smaller companies improved relative to mega-cap technology toward month-end.

 

Outlook We maintain a relatively positive market outlook. The global economy remains robust despite inflationary headwinds and higher interest rates. Corporate earnings growth continued in the third quarter and looks healthy for the full year. AI infrastructure demand remains the top earnings driver. However, valuations are elevated and profitability concerns persist. Financing issues have also arisen. A broadening out of market performance is likely. This comes as investors continue to diversify away from the AI theme. Overall, our approach remains unchanged. We focus on company fundamentals and long-term growth potential. This discipline helps us seek opportunities while managing risks in a market shaped by rates, AI spending and uneven growth.