Electric Dreams: How AI is Driving a Generational Opportunity in Electricity

Electric Dreams: How AI is Driving a Generational Opportunity in Electricity

Cleantech equities have broadly rebounded, and there are signs of renewed investor interest. 

Traditionally, electricity has been one of the more stable—some might say dull—segments of the energy sector. For many years, global electricity demand grew at a rate of one percent a year, a modest pace that reflected mature usage patterns and efficiency. But that is all changing, very quickly. In fact, we are witnessing a seismic shift in electricity demand globally, driven by transformative technological and industrial trends. Today, the emergence of large-scale Artificial Intelligence (AI) data centres, rapid electric vehicle (EV) adoption, intensified heating and cooling needs, and the reshoring of energy-intensive manufacturing are ushering in a new era of secular electricity growth.

For electricity demand, we believe this is a generational inflection point, and at its core is the boom in AI data centres. AI workloads, particularly training and inference for large language models, consume significantly more electricity than traditional enterprise workloads. According to BloombergNEF, global data-centre power demand is expected to rise from around 449 terawatt hours in 2025 to about 1,600 TWh by 2035, meaning average hourly electricity use will nearly triple.

To put that in perspective, total global electricity consumption across all sectors reached 26,600 TWh in 2024 and is projected to grow to 35,600 TWh by 2035. Data centres alone are expected to account for 14% of that increase, expanding their share of global electricity use from 1.4% in 2024 to 4.5% by 2035. By comparison, the Hoover Dam, one of the most iconic hydropower plants in the United States, has a generating capacity of 2,080 megawatts and produces about 4.5 TWh annually. The projected annual increase in global data centre demand is therefore equivalent to the output of roughly 25 Hoover Dams.

These figures highlight the industrial-scale impact of AI data centres on electric grids globally. Their unique load profiles, high baseline demand, large batch-training loads, intensive cooling requirements and geographic clustering are already straining grids. Large “hyper-scaler” tech companies are securing long-term power purchase agreements, often at premiums, to ensure access to 24/7 clean power.

Beyond AI, other factors are contributing to rising electricity demand. Rapid EV adoption and the reshoring of manufacturing, including battery plants and semiconductor fabrication plants, are expected to significantly boost power demand. EVs alone could add 6%, or 1,610 TWh, to global electricity consumption by 2035. Additionally, longer hot seasons and expanded heating/cooling deployment are stretching peak demand and stressing grid capacity. The broader load-growth trend is systemic and likely to play out over decades.

These converging trends are reshaping power-system economics and investment priorities. We see the following as key implications:

• Firm, flexible power is essential: Predictable and dispatchable generation such as nuclear (especially small modular reactors (SMRs)) and geothermal, as well as long-duration storage, are going to be critical.
• Storage and flexibility: Batteries and storage systems will be key to supporting 24/7 AI workloads and fast-changing demand, which will go far beyond just shifting solar and wind output.
• Grid modernization: The current grid is designed for slow growth. Now, it requires significant upgrades to accommodate new loads and variable renewables.
• Speed-to-power: Traditional generation build cycles are too slow. Efficient project delivery within 12 to 36 months is now a strategic advantage, favouring gas turbines, battery storage and modular solutions.

The emerging electricity demand landscape is rife with opportunities, but there are risks and challenges to meet surging power demand. In particular, the grid remains a vulnerable link in the energy transition. Lengthy permitting processes due to regulatory hurdles, skilled workforce gaps in engineering and construction, and supply-chain constraints for critical electrical equipment components pose significant risks to timely execution and deployment of new energy projects.

Much has been made in recent months of the U.S. retreat from climate-action leadership at the federal level, but the U.S. cleantech sector continues to show resilience and solid growth. States like Texas and Oklahoma, historically fossil-fuel strongholds, are now leading in wind and solar deployment, underscoring how economics, not policy alone, are driving investment decisions.

Clean Energy Stocks Rebounded in 2025

Source: Bloomberg LP as of November 12, 2025. Past performance is not indicative of future results. One cannot invest directly in an index.

Cleantech equities have broadly rebounded and are outperforming the broader market year-to-date, correcting the previous excessive discounting that arose from policy uncertainty and macro headwinds, and there are signs of renewed investor interest. Capital flows into clean energy markets have picked up, with investors increasingly favouring companies exposed to grid modernization, energy storage and dispatchable clean power.

Valuations in select subsectors, particularly long-duration storage and nuclear SMRs, are beginning to re-rate, as investors recognize the strategic importance of firm, flexible power in supporting AI-driven demand. State-level mandates and certain incentives in the U.S. Inflation Reduction Act continue to provide tailwinds, even as federal momentum slows.

The resurgence in investor interest in cleantech energy points to the reality that the AI revolution is not only a technological shift, but also an energy transformation. As data centres proliferate and electrification accelerates, the grid must evolve at an unprecedented pace. Success will hinge on delivering clean, firm and fast electricity that aligns innovation with sustainability, reliability and scalability. For investors, the convergence of AI-driven demand and the energy transition may well present a once-in-a-generation investment opportunity.