Residential heat is 13% of global energy. So what if AI could optimize residential electric heating? This 16-page report finds that AI can load-shift hot water tanks, unlocking 1.5-8.5% additional flexibility in grids, and making air source heat pumps the lowest cost option for heat in Europe, eclipsing gas-combi boilers, saving $300 per household per year?
More flexible grids could help resolve bottlenecks as global electricity demand grows by 4% pa to 2050. The key idea is that today’s existing grids have c15-50% spare capacity to support additional loads, as long as those loads scale back during the 1-10% most strained hours, per pages 2-3. Some economic benefits of load-shifting are quantified on pages 3-4.
Hence in this report, we want to explore if AI can load-shift hot water tanks. I.e., can the hot water tanks in residential and commercial heat be used for load shifting, smoothing the grid by absorbing excess electricity during times of grid surplus and cheap power pricing, making hot water, storing it, and then avoiding using electricity during times of grid strain and high pricing?
This hinges on homes having electric heating, preferably heat pumps, smart meters and time-of-use tariffs. The size of electric heating in both the US and Europe is discussed on pages 5-6, including the rise of heat pumps, on page 7, and the rise of time-of-use tariffs on page 8.
We built our own optimizer, for the heating needs of a medium-large house in Northern Europe with a ground-source heat pump, and back-tested the optimizer, hour by hour, using data from 2025. Using AI to load-shift a hot water tank can reduce the wholesale electricity costs of residential heating by 50%, per pages 9-11.
We modeled the costs of different home heating systems, such as gas, oil, immersion heating and heat pumps, in both the US and Europe. Fascinatingly, in Northern Europe, an AI optimized heat pump becomes the lowest cost option, even beating a gas combi boiler. This is specifically because of the cost deflation offered by load-shifting the heat pump to times of low electricity prices, per pages 12-13.
The benefits of load-shifting could be enhanced further by expanding the scope of time-of-use tariffs, or the structure of time-of-use tariffs, to include T&D incentives as well as just wholesale power market incentives. More on this observation on page 14.
Leading companies that offer load-shiftable heat pumps, load-shiftable hot water tanks, or smart optimization algorithms are discussed on pages 15-16. This theme really fascinates us, as another example where technologies collide.