AI energy demands threaten to offset climate efficiency gains
A Bank for International Settlements paper finds that artificial intelligence poses dual climate effects, balancing productivity gains against rising electricity demand. The net impact on inflation and financial stability depends on infrastructure investment and technological scenarios.
Efficiency gains meet power constraints
The study evaluates how artificial intelligence reshapes the climate transition through opposing mechanisms.
Machine learning tools improve resource optimization, accelerate green innovation, and refine risk forecasting.
However, training and running large models require heavy computational infrastructure, driving up electricity consumption and associated carbon emissions.
Net climate outcomes remain uncertain due to rebound effects where efficiency gains spur higher overall usage.
Under an AI copilot scenario, productivity benefits and environmental costs develop gradually over time.
In contrast, artificial general intelligence introduces much greater technological disruption, magnifying both mitigation capabilities and environmental hazards across global power systems.
Macroeconomic stakes for monetary policy
The AI-climate nexus directly affects central bank mandates through several economic channels.
Shifts in energy systems, productivity, and climate hazards alter potential output and inflation dynamics.
Massive capital requirements for computing facilities and power grids also create distinct financial stability considerations.
Central banks must monitor whether large-scale debt financing for digital infrastructure elevates credit risks or leaves lenders vulnerable to rapid technological obsolescence.
Abstract scenarios over concrete risks
The analysis rightly flags the computing trade-offs of artificial intelligence, but relies excessively on speculative scenarios.
Central banks risk diluting their core mandates by treating hypothetical tech trajectories as near-term policy inputs.
Prudence requires tracking real balance sheet exposures rather than policing energy consumption.