Artificial intelligence is often criticized for its massive energy demands, particularly as data centres expand to handle the technology’s rapid growth. Yet experts argue that AI’s long-term potential to improve efficiency across industries could make its climate benefits outweigh its costs.
Rising Energy Demand
The immediate challenge is clear: powering AI requires more electricity. To meet surging demand, some companies are investing in gas-fired power plants, while others — like Google — are exploring future-oriented options such as nuclear fusion. Renewable energy alone struggles to guarantee the reliability needed for energy-hungry data centres.
Today, data centres consume about 415 terawatt hours (TWh) of electricity annually, roughly 1.5 per cent of global demand, according to the International Energy Agency (IEA). By 2030, this figure could more than double to 945 TWh. However, compared with the additional demand expected from electric vehicles and air conditioning, data centres will remain a smaller piece of the overall energy puzzle.
Efficiency Gains Across the System
The bigger story may be AI’s role in cutting energy waste. The technology is expected to enhance efficiency across industries, potentially saving far more energy than it consumes. For climate purposes, reducing emissions today is more urgent than tomorrow, but the potential scale of AI-enabled efficiency gains makes the trade-off significant.
Some sectors are especially inefficient. The materials value chain — producing steel, glass, hydrogen, ammonia, copper and other essentials — consumes four to five times more energy than theoretically required, according to consultancy Thunder Said Energy. AI could help close that gap by discovering new materials, catalysts, or processes, in much the same way it has already accelerated breakthroughs in biotechnology.
AI in Battery Innovation
One promising area is battery technology. AI is being used to screen millions of potential materials for solid-state batteries, which could enable longer-lasting, lighter, and more efficient energy storage. Microsoft, working with a U.S. government lab, has already narrowed down tens of millions of candidates to just 23 viable electrolyte materials for lithium-based batteries. Such efforts highlight AI’s ability to accelerate research that might otherwise take decades.
Smarter Energy Use
Beyond research labs, AI could help reduce waste in everyday energy use. Smart appliances, connected sensors, and optimized grids are expected to play a growing role in cutting inefficiencies in production, transport, and consumption. Even modest savings across such large systems could translate into significant reductions in electricity demand.
Net Positive for the Energy Transition
While AI-driven data centres add to global electricity use, the broader application of AI may ultimately prove a net positive for the energy transition. By tackling inefficiencies in materials production, battery development, and energy management, AI has the potential not just to balance its own footprint but to advance progress toward global climate goals.
