How AI-driven energy demand is creating large-cap growth investment opportunities
A prominent tech CEO recently compared artificial intelligence to the mastery of fire or electricity in terms of its importance to humanity. While bold, this assertion carries weight when considering AI’s transformative potential for economic and social conditions.
Fire enabled human evolution by transforming our ancestors’ diet and brain development. Electricity has driven the modern global economy for over a century, with its adoption directly correlating to societal development across nearly every metric.
If AI achieves even a fraction of its potential, access to energy will be vital in making that reality possible.
This creates a compelling investment opportunity beyond chip manufacturers and AI companies: power generation infrastructure.
AI energy surge
The story begins with hyperscalers — Alphabet, Amazon, Meta Platforms, and Microsoft. Moody’s estimates these companies spent over $210bn on capital expenditures in 2024, largely for AI infrastructure. State Street forecasts this spending will exceed $320bn in 2025.
Power availability is becoming a critical bottleneck for AI development. AI data centres require significantly more electricity than traditional ones due to specialized chips that generate more heat and demand more power.
According to Goldman Sachs, a ChatGPT query requires approximately 2.9 watts — almost 10 times the electricity of a traditional Google search.
McKinsey predicts data centres could increase their electricity consumption from 3-4% of total US electricity today to 11-12% by 2030 — jumping from 25 gigawatts to 80 gigawatts.
US’s energy infrastructure challenge
This surge in AI energy demand coincides with broader strains on US energy infrastructure. After decades of declining electricity demand due to deindustrialisation and improved energy efficiency, the situation has reversed. We’re experiencing a power supply crunch from years of underinvestment, coinciding with surging demand from AI industries and manufacturing reshoring incentives.
The infrastructure shows its age
According to Merrill Lynch analysis, nearly one-third of US transmission and almost half of distribution infrastructure is nearing end-of-life. Even newer power generation capacity built in the early 2000s now requires refurbishment, as gas turbines have 20–30-year replacement cycles.
Energy solutions in focus
There are four key areas available to meet surging energy demand: natural gas, nuclear power, renewable energy, and greater efficiency.
- Natural gas: The growing need for reliable baseload power drives natural gas demand. European countries seeking alternatives to Russian gas have turned to the US for supply. The Trump administration’s “National Energy Emergency” executive order has eased permitting and regulatory concerns, facilitating fossil fuel infrastructure buildout. GE Vernova exemplifies this opportunity. As the leading natural gas turbine manufacturer, the company’s installed base accounts for 25% of power produced worldwide and over 55% of U.S. natural gas power generation. The company maintains a growing backlog of orders stretching ahead for several years.
- Nuclear power: Nuclear offers reliable, clean, around-the-clock energy, spurring efforts to restart inactive reactors and approve small modular reactors (SMRs). SMRs have about one-third the capacity of traditional reactors but are cheaper and quicker to construct, with better safety profiles. However, nuclear investments remain highly speculative. Companies such as NuScale Power generate no revenue while incurring massive capital expenditures. While NuScale recently announced a deal with Tennessee Valley Authority, the company previously ended an earlier contract due to higher-than-expected costs. These projects carry unique risks and require years to materialize.
- Renewables: Despite Trump administration opposition, market forces continue driving renewable investment due to relatively lower costs and faster deployment. However, regulatory uncertainty creates hurdles. The administration cancelled two advanced wind projects in 2025, questioning the viability of such initiatives.
- Energy efficiency: Paradoxically, AI technology causing power concerns also offers solutions. AI could improve energy efficiency and power company operations by making distribution systems more reliable and aiding fuel source discovery. Continued efficiency improvements in consumer goods, industrial processes, and transportation technology point toward greater efficiency across economic sectors.
Investment risks and strategy
The persistence of AI-driven data center demand represents the key risk to this investment thesis. While we believe this trend is durable, many question whether such substantial AI capital expenditures can last. These concerns echo the early 2000s tech boom energy infrastructure overbuilding, which resulted in years of decreased power generation project demand.
The power mix may also differ from expectations. Political and regulatory environments significantly impact renewable energy viability, while nuclear projects face cost and timeline uncertainties.
In addition, power generation and distribution projects are expensive, often taking years to complete while carrying regulatory risks.
This argues for a “picks-and-shovels” investment approach through companies like GE Vernova, which produces turbines and components for wind, nuclear, and natural gas plants.
This strategy benefits from energy infrastructure spending while remaining agnostic about specific power generation forms.
Conclusion
AI’s comparison to fire and electricity may prove far-sighted.
As AI transforms society, the companies powering this revolution present compelling investment opportunities. However, thoughtful company-specific analysis remains essential. Just as fire requires proper precautions, investing in energy companies demands careful consideration and understanding broader context.
The structural trends favour investment in power generation infrastructure, but success requires thorough security selection and recognition that this is a probabilistic exercise involving various possible outcomes.
For investors willing to conduct diligent research, the AI energy revolution offers significant potential returns.