Powering the AI Revolution: NVIDIA's Data Center Dominance and the Energy Infrastructure Play

The rise of artificial intelligence has reshaped the global economy, and at the heart of this transformation lies NVIDIA. With its dominance in data center GPUs—commanding an estimated 80–90% market share—NVIDIA's hardware underpins everything from generative AI models to autonomous systems. Yet, this revolution comes at a cost: energy consumption. Data centers powered by NVIDIA's GPUs now demand exponential amounts of electricity, creating a critical opportunity for energy infrastructure firms positioned to meet this surge.

NVIDIA's Data Center Dominance: A Powerhouse with a Price

NVIDIA's Blackwell GPU architecture, introduced in 2025, exemplifies the trade-off between computational power and energy use. These chips deliver a 40× performance leap over predecessors but consume up to 1,200 watts, with paired systems (e.g., the GB200) drawing 2,700 watts—a 300% increase over earlier generations. This has driven NVIDIA's data center revenue to $35.6 billion in Q4 2024, a 93% year-over-year surge.

Yet, the energy required to fuel these GPUs is staggering. Analysts estimate that global data center power demand could triple by 2027, exceeding 2022's total energy consumption if utilization rates hit 90%. This creates a paradox: AI's growth hinges on infrastructure that can scale energy supply while mitigating costs and emissions.

The Energy Infrastructure Play: Undervalued Stocks to Watch

The companies best positioned to capitalize on this demand are not silicon foundries but energy infrastructure firms. Five undervalued stocks identified by Morningstar analysts—Chart Industries (GTLS), Energy Transfer LP (ET), Enbridge (ENB), Kinder Morgan (KMI), and TC Energy (TRP)—offer exposure to critical energy systems supporting data centers. Here's why:

1. Chart Industries (GTLS): The Carbon Capture Specialist

GTLS designs equipment for carbon capture, gas storage, and specialty compression. With a $3.35 billion order backlog (double its 2022 total), GTLS is supplying the “batteries” for data centers: systems that store energy and reduce emissions. While its $200 fair value estimate (vs. current price) reflects growth potential, risks include execution challenges from recent acquisitions.

2. Energy Transfer LP (ET): Fueling Texas's Data Hub

ET's pipelines connect to 55–60% of Texas power plants, a key data center cluster. The firm projects 8 billion cubic feet/day of gas demand by 2030 from data centers alone. With an 8.23% dividend yield and a $21 fair value estimate, ET offers income and growth.

3. Enbridge (ENB): The Northern Grid Guardian

Enbridge's gas storage and renewables infrastructure serve regions like Ohio and North Carolina, where 45% of North American power generation lies within 50 miles of its assets. Its $41 fair value estimate and 7.54% dividend yield make it a stable play on data center energy needs.

4. Kinder Morgan (KMI): The Gas Transport Titan

KMI controls 20% of U.S. power demand and 40% of gas transport infrastructure. With AI-driven gas demand projected to hit 16 billion cubic feet/day by 2030, KMI's pipelines are indispensable. Its $22 fair value estimate and 6.08% dividend yield offer value despite a 24.98 P/E ratio, elevated but justified by its scale.

5. TC Energy (TRP): The Storage Solution

TRP's 650 billion cubic feet of gas storage (80% under take-or-pay contracts) acts as a “battery” for peak data center demand. Its $47 fair value estimate and 7.44% dividend yield highlight its reliability. However, its 5.36 P/B ratio underscores valuation risks.

Risks and Considerations

The path to profit is not without hurdles. Supply chain bottlenecks (e.g., steel and copper shortages), regulatory delays, and labor constraints could slow infrastructure expansion. For instance, U.S. grid interconnection timelines can stretch to 7 years, threatening data center timelines.

Moreover, the energy transition itself poses risks. While gas remains critical for grid stability, renewables' intermittency could amplify demand for gas “backstops.” Companies like GTLS and ET must balance these dynamics.

Investment Thesis: A Strategic Bet on Energy Resilience

Investors seeking to profit from AI's energy demands should prioritize dividend-paying infrastructure stocks with low valuations relative to their fair value. The five firms above offer exposure to gas transport, carbon capture, and storage—critical for data center reliability.

• For income investors: ET, ENB, and TRP's high dividends (7–8%) provide steady returns while infrastructure grows.
• For growth investors: GTLS and KMI's order backlogs and scalability position them to outperform as AI adoption peaks.

Avoid overpaying: while KMI's P/E is high by sector standards, its historical 10-year average (67.31) suggests current valuations are reasonable.

Final Word: Powering Progress, Prudently

NVIDIA's AI revolution is undeniably transformative, but its success hinges on energy infrastructure that is robust, scalable, and sustainable. The companies highlighted here are not just beneficiaries of this trend—they are enablers. For investors willing to navigate near-term risks, these undervalued energy stocks offer a compelling way to profit from the next era of computing.

In the race to power AI, the winners will be those who supply the electricity—and the wisdom—to keep the servers running.