Will AI help or hinder the energy transition?
Key takeaways
AI has the potential to accelerate the energy transition by improving efficiency, optimizing power systems, and supporting industrial decarbonization, while increasing electricity demand creates new challenges.
- AI is expected to increase electricity demand, with data centers projected to account for 3% of global electricity usage by 2030, even as AI applications reduce energy use and improve operational efficiency.
- AI helps utilities improve grid operations through demand and renewable generation forecasting, predictive maintenance, dynamic line rating, and optimization of distributed energy resources and battery storage.
- AI also supports industrial decarbonization, climate resilience, and energy efficiency improvements, which remain essential to achieving net zero goals.
Artificial intelligence (AI) is driving a surge in electricity demand. Yet it is also helping utilities and industrial users improve efficiency, reduce emissions and accelerate decarbonization.
AI technology has caught investors’ attention for its impact on efficiency and productivity. The implications for the energy transition, however, are a topic of debate.
While AI’s efficiency gains have the potential to accelerate decarbonization, the big question is whether they will outweigh AI’s own considerable energy footprint. Data centers are projected to account for 3% of global electricity usage by 2030 as AI adoption ramps up, with fossil fuels a key source of generation. (See Figure 1.)
Andrey Berdichevskiy, Partner and Associate Director at Boston Consulting Group (BCG), who leads the firm’s Southeast Asia Center for Climate & Sustainability, pointed out that AI’s many applications in optimizing scarce resources and financial returns also serve to improve sustainability outcomes.
Berdichevskiy and his colleagues calculate that AI-enabled sustainability applications could createhundreds of billions of dollars of annual value over the coming decade through lower energy use, improved operational efficiency, and reduced capital requirements. (See Figure 2).
Across the sectors analyzed by BCG, including electricity and industry, the evidence suggests that AI-enabled optimization would lead to fewer resources used in aggregate.
As for whether AI is ultimately a net positive or negative for the energy transition, the International Energy Agency (IEA) noted that while concerns that AI could increase emissions appear overstated, so too are hopes that it could singlehandedly address the issue.
The IEA highlighted that there remain significant uncertainties about the outlook for AI-related electricity demand, which will have an impact on the rate of adoption and the development of more energy-efficient models. Meanwhile, it said that the energy sector could go a lot further in harnessing AI’s potential to enhance efficiency.
Managing a more complex power system
While the rapid expansion of renewable energy deployment over the past decade has reduced emissions from electricity generation, it has also introduced new operational challenges.
“Managing that complexity is where AI is coming in,” said Graeme West, Professor of Electronic and Electrical Engineering at the University of Strathclyde and Critical Infrastructure Theme Lead at the Alan Turing Institute.
West identified four areas where AI is already having a significant impact: grid operations, asset management, grid planning and electricity markets.
On the operational side, AI can improve forecasting of both electricity demand and renewable generation, helping system operators balance supply and demand more efficiently. On the asset management side, AI supports predictive maintenance.
“If you understand the current health and future health of the system, then you can optimize where you go and invest and upgrade the grid,” said West.
Utilities face growing pressure to adapt to changing supply-demand dynamics due to the growth of electric vehicles, heat pumps, renewable generation and energy-intensive data centers. Machine-learning models can help identify where networks require reinforcement and how existing infrastructure can be used more effectively.
One example, noted West, is dynamic line rating. Transmission lines are traditionally assigned conservative capacity limits based on assumed weather conditions. By combining sensors, forecasting and machine learning, operators can determine in real time how much electricity a line can safely carry, potentially unlocking additional capacity without building new infrastructure. This could prove especially valuable a time when grid investment is struggling to keep pace with rising demand.
Emily Kirsch, Managing Partner at Powerhouse Ventures, an early-stage investor in software solutions to support grids and clean energy, noted that flexibility is becoming one of the defining themes of the modern electricity system.
“The grid is becoming harder to manage,” she said. “More inverter-based resources, variable generation, extreme weather events and an aging workforce all compound the challenge. AI is becoming essential to operating it reliably.”
The rise of distributed energy resources is creating additional opportunities. AI can coordinate electric vehicle charging, behind-the-meter batteries, rooftop solar systems and flexible demand to support grid stability.
West pointed to the growing adoption of smart tariffs as a case in point. Rather than consumers deciding when to charge their vehicles, AI increasingly determines the optimal charging times based on electricity prices, renewable energy availability and grid conditions.
Storage is becoming an especially important part of this equation. As battery deployment accelerates, AI is increasingly being used to optimize dispatch decisions – in other words, when to charge, when to discharge, and at what power level.
Meanwhile, for operators of battery energy storage systems (BESS), sophisticated AI-driven forecasting models can help maximize revenues while supporting system reliability. Without that analytical layer, Kirsch said, “the economics of battery storage become harder to underwrite.”
Improving industrial efficiency
While electricity and heat production accounts for roughly a third of global greenhouse gas emissions, industrial activities contribute nearly a quarter.
Using AI to optimize the use of resources will support industrial decarbonization at the same time as improving financial returns, said Berdichevskiy.
“If you improve the efficiency of your machinery, you’re saving costs in the form of energy, but you’re also using less energy, which produces fewer emissions,” he said.
One of AI’s greatest contributions to decarbonizing industry could be in energy-intensive sectors such as cement production, chemicals manufacturing and steelmaking.
West highlighted London-based AI company Gigaton, which uses AI to improve process efficiency in these hard-to-abate industries. “If you do things more efficiently, you take huge amounts of carbon out of the process,” he said.
Bolstering climate resilience
AI can also improve climate resilience, particularly in sectors exposed to extreme weather and natural disasters, said Berdichevskiy.
He pointed to Neara, a company that creates digital twins of electricity networks. By combining engineering models, weather data and AI-powered analytics, the platform can identify sections of transmission infrastructure that may be vulnerable to wildfire risk.
“Previously, utilities had to inspect these assets manually,” said Berdichevskiy. “Now you can model the entire network and identify where intervention is needed before problems occur.”
Better risk assessment can also improve insurance availability and pricing, making infrastructure assets more attractive to investors.
Reframing energy transition investment
To date, climate investment has focused on hardware: renewables, electric vehicles, batteries, and other forms of low-carbon infrastructure. Yet AI and other solutions that improve energy efficiency are often excluded.
“There is a need to reframe what counts as climate investment or energy transition investment,” Berdichevskiy said.
That is, after all, a key focus for policymakers. Achieving an annual average 4% rate of energy efficiency improvements is a critical pillar of the International Energy Agency’s net zero by 2050 scenario – a target that was adopted by 130 countries at the 2023 UN Climate Change Conference in Dubai.
Given that efficiency gains stood at 1.8% in 2025, a large gap remains. Going forward, AI will no doubt play an increasingly important role in closing it.
To learn more about energy transition investment trends, explore these stories:
How new tech is driving new investment in nuclear power
Grid and battery investment: Why it matters to the energy transition
Why data centers are driving investment in alternative energy
The long boom in data infrastructure
You may also be interested in
Explore our programs and certificates
CFA Institute offers a diverse range of programs and certificates designed to meet the needs of finance professionals across various career stages and specializations.
