Disentangling the evolving effects of climate change and technology diffusion on building electricity demand

Climate change and technology diffusion will jointly drive future growth in building electricity demand. How this nonadditive response evolves over time with the pace of climate change and technology diffusion has not previously been examined, despite its importance for maintaining reliable power systems. We estimate electricity demand from residential and commercial buildings under climate change and/or technology diffusion across 183 counties in the United States through 2060. Our framework drives 209,352 machine learning models of building electricity demand with 100 spatiotemporally downscaled climate realizations across four climate models and three emissions scenarios, and 18 technology diffusion pathways of air source heat pumps and insulation. We find that the combined response of peak demand to climate change and technology diffusion diverges substantially from the sum of their individual responses, with distinct patterns across hot and cold grid regions. In cold regions, climate change and technology diffusion drive roughly up to 10 times larger peak demand growth than climate change alone by 2060. Conversely, in hot regions, technology diffusion counteracts climate-change-driven peak demand growth. Across regions, climate-change-induced warming reduces or eliminates a shift toward winter peaking driven by technology diffusion. Via uncertainty partitioning, we find climate-related uncertainty dominates near-term peak demand uncertainty, while technology-diffusion-related uncertainty increases to 10 to 80% of uncertainty by 2060. Internal variability accounts for a much larger share of climate-related uncertainty in peak demand than in temperature itself. Our results demonstrate that capturing nonadditive responses between climate change and technology diffusion is important for planning reliable power systems.