Battery owners usually begin with the same question: how much can the system earn? Our experience in Lithuania shows why there is no one-size-fits-all solution.
Four months of measured data from an industrial site demonstrate the value created by optimizing a battery against day-ahead electricity prices. The result represents savings on the site’s electricity bill, not measured income from balancing services or wholesale trading.
Battery performance depends on more than power and capacity. Site consumption, solar generation, the electricity contract, network charges, export rights, and the control strategy can all affect the result.
Measured results
The clearest example in our current dataset is a Lithuanian industrial site with its own solar installation. Its battery has approximately 1 MWh of capacity and 0.5 MW of power, making it roughly a two-hour system.
The site buys electricity at the Nord Pool day-ahead price for Lithuania, plus a supplier margin and network charge. The site was not paid for electricity exported to the grid during the period analyzed.
From April to July 2026, our software controlled the battery according to day-ahead prices. It created an average value of approximately €1,960 per installed MWh of capacity per month. Monthly results ranged from €1,640 to €2,230 per installed MWh.
The battery charged when electricity was cheaper or surplus solar energy was available. It discharged when the site would otherwise have bought more expensive electricity.
Around 43% of the charging energy came from the solar installation and 57% from the grid. The battery completed approximately 0.8 equivalent full cycles per day. In July, the average electricity price during charging was €0.070/kWh, compared with €0.145/kWh during discharge – a difference of around €75 per MWh.
What the figures do – and do not – show
Multiplying the four-month average by 12 gives an indicative value of approximately €23,500 per installed MWh per year. This is a simple extrapolation, not a measured annual result, as the data cover only spring and summer.
The calculation also excludes potential revenues from intraday trading and balancing services. Although our platform can exchange data for FCR, aFRR, and mFRR offers, the batteries in this dataset have not yet participated in these markets. Thus, we cannot provide a measured revenue split between them for now.
A different site illustrates how site conditions and operating strategy can affect the result. There, a similarly sized battery was used mainly to increase onsite solar consumption while operating under a grid export constraint. It created approximately €460 per installed MWh in July – roughly one-quarter of the first example’s monthly revenues.
Whether exported electricity is paid for, how much export capacity is permitted and how network tariffs are structured can influence the business case as much as the battery hardware.
Across the batteries we have measured, discharged energy amounted to between 93% and 97% of charged energy. These efficiency losses are already included in the results. Battery degradation, financing costs, fixed network charges and possible trading or aggregation fees, however, are not. The figures should therefore not be read as net profit.
The schedule still has to reach the battery
An optimization model has little value if the battery does not execute its instructions correctly and reliably. The energy management system must combine price data with forecasts of consumption and solar generation, monitor state of charge, respect operating limits and send instructions that the site controller can execute safely.
In the systems we manage, the controller normally confirms a command within 0.5 to two seconds. The actual power response takes longer: the median is 39 seconds, while the slowest 10% of experienced responses take up to 11 minutes. Most of this delay comes from the battery and inverter, with significant differences between manufacturers.
(Additional note on delays. In our data, the slow cases cluster around state transitions, especially waking from idle to an active power level or returning to idle; the worst case was a unit that took up to 15 minutes to respond to an idle command. The same battery can respond in 30 seconds or in several minutes within the same hour depending on the transition requested, and the behavior was observed across several manufacturers and sites, with large differences between them: the median response ranged from about 10 to about 45 seconds depending on the manufacturer. This is why we treat response time as a property of the battery model and the requested transition, not of the control platform, and measure it per battery model before committing to fast-responding services.)
A lost connection, an outdated schedule or a controller that fails to follow commands can reduce revenue and, when participating in electricity markets, create imbalance costs.
Accurate metering is equally important. We have encountered errors that caused the calculated battery value to change from negative to significantly positive once the data had been corrected.
Aggregation is particularly relevant for commercial and industrial (C&I) batteries. In Lithuania, balancing offers start at 1 MW and use 1 MW increments, while commercial batteries are commonly rated at only 0.1–0.5 MW. Most individual assets therefore need to join a larger portfolio before they can participate.
For an owner, the starting point is accurate metering, a tested control connection, and an electricity contract suited to how the battery will be used. The battery’s response and behind-the-meter value should be demonstrated before it is aggregated and qualified for balancing services.
Our data show why a revenue estimate should begin with the conditions at the individual site, rather than a generic figure for the battery.
About the author
Robertas Janickas is the CTO at Inion Software. He has over 10 years of experience in management and software and hardware development. Inion Software develops a single platform for monitoring solar and wind power plants and managing battery storage and energy flows.
The post How much can a battery actually earn? Four months of data from Lithuania appeared first on Energy Storage.
