Home batteries change the cheap electricity prices they chase

Home batteries can save money by charging when electricity is cheap and discharging when it is expensive. The worry is that identical price signals make them move together, creating a new overnight demand peak on the transformer serving a neighborhood.

A preprint by energy economist Lion Hirth argues that many simulations leave out an important response: widespread charging changes the wholesale prices that encouraged it. In his German-calibrated model, including that feedback turns a large increase in neighborhood import peaks into a modest reduction. This is a simulation result, not evidence that real grids are automatically protected.

The price is part of the response

An individual household is too small to move national electricity prices appreciably. A large fleet of batteries is not. Charging adds demand during cheap hours, pushing prices upward; discharging adds supply during expensive hours, pushing prices downward. Both actions narrow the price differences that make buying low and selling high attractive.

A simulation that holds prices fixed as batteries proliferate misses this collective effect. Hirth reports that none of the home-battery studies in his review included wholesale price feedback.

Feedback need not mean changing a household’s tariff after it has charged. Retailers can anticipate customers’ responses and incorporate that expected demand into their wholesale-market bids. The model assumes those expectations are correct and calculates prices and battery operation together.

Two versions of the same neighborhood

Hirth simulated a year of hourly electricity flows for 100 single-family households, half with rooftop solar. Battery ownership increased until every solar household had one: 50 batteries across the neighborhood, not batteries in every home. Each battery had 8 kilowatt-hours of capacity and could charge or discharge at 4 kilowatts.

Households minimized their electricity bills using advance knowledge of prices, consumption and solar generation within each modeled month. Their tariff combined wholesale prices with fixed network charges and taxes. Charges were refunded on grid electricity subsequently returned to the grid, apart from losses.

For the national calculation, the neighborhood was replicated across 190,000 identical grids. Wholesale prices responded to national demand after accounting for wind, solar and storage. The price relationship was calibrated to German data for 2025; other flexible resources were represented collectively as utility-scale storage.

The main grid measure was the average import load during the highest 1% of hours, rather than the single largest hourly demand. At maximum battery adoption, fixed wholesale prices made this measure 83% higher than without home batteries. With price feedback, the same measure was 6% lower. The corresponding export measure was 15% below its no-battery benchmark with feedback.

Charging became less concentrated, and narrower price spreads made buying electricity for later resale less attractive. Batteries consequently devoted more of their operation to storing household solar production. Because homes differed in consumption and solar output, that also reduced synchronization.

National balance is not local protection

A simplified analytical model explains the geographical catch. When charging-power limits do not bind, the local charging peak depends on the neighborhood’s share of national flexibility, the energy absorbed nationally and the shape of the national demand trough after wind and solar generation are subtracted.

A neighborhood with unusually many batteries can still create a local peak without moving national prices enough to prevent it. In a sensitivity run with the same local battery adoption spread across only a quarter as many grids nationally, the import measure remained 24% above the no-battery benchmark despite price feedback.

Alternative price relationships and household assumptions generally preserved the reduction relative to fixed-price simulations. They did not establish that peaks always disappear. Larger batteries and measured household consumption profiles produced increases in some cases. Even in the base case, the single highest hourly import was 3% above its no-battery counterpart with feedback, illustrating why the definition of a peak matters.

An equilibrium still needs implementation

The paper identifies another consequential limitation: calculating a balanced outcome does not prove households will implement it. When identical batteries face equal prices across several hours, their owners may be indifferent about when to charge. Prices alone need not select the smooth charging schedule calculated by the model. A slightly cheaper hour could instead attract everyone together.

The simulations assume the calculated schedule is implemented. Direct retailer control under flexibility contracts can provide a way to do that; independent household responses to announced prices are less straightforward.

The study measures aggregate loading, not a detailed engineering assessment of every cable, voltage constraint or reinforcement cost. Its code and inputs were available on request, with public release planned upon acceptance. The practical conclusion is therefore narrower than reassurance: grid planners should include wholesale price feedback, while still checking local concentration and whether the assumed charging schedules can actually happen.