For years, the business case for energy storage at commercial and industrial facilities has largely been built around what happens when the grid is working: reducing demand charges, shifting consumption away from expensive hours, or participating in demand-response programs. The cost of what happens when the grid is not working has been harder to fit into that same calculation, and new data from Oak Ridge National Laboratory makes the omission increasingly difficult to defend.
What the Outage Data Shows
Researchers at Oak Ridge National Laboratory found that major U.S. power outages cost customers an average of more than $67 billion annually between 2018 and 2024, with the total reaching $121 billion in 2024 alone. Commercial and industrial customers absorbed an average of $6,031 per outage in 2024, and businesses in every state saw average per-outage costs above $4,000. Between 2018 and 2024, the number of major U.S. outages rose 29%, from 4,666 to 6,533, and the average outage lengthened from 9.6 hours to 11.8 hours. In one South Carolina county the lab examined, the estimated loss reached $140,000 per non-residential customer for a single outage.
At the same time, the technology companies can use to offset some of that exposure has become considerably cheaper. BloombergNEF found that battery packs for stationary storage averaged $70 per kilowatt-hour in 2025, down 45% from 2024 and the steepest decline of any battery segment the firm tracks, making stationary storage the least expensive lithium-ion category for the first time. Lithium-ion pack prices across all applications averaged $108 per kilowatt-hour. Rising outage costs and falling storage prices are, in effect, two lines on the same graph moving in opposite directions, and that shift changes the shape of the storage conversation.
The ROI Calculation Has Been Missing a Cost
An energy storage project evaluated only on utility savings can look very different from one evaluated against the financial consequences of losing power. A $6,031 average outage does not by itself justify a battery, but the cost belongs in the model, alongside the broader tradeoffs resilience spending already faces against other capital priorities.
For businesses, ORNL calculates outage costs using lost production and direct interruption costs, offset by any energy savings realized during the outage itself. That makes the data particularly relevant to finance teams, because the loss is not an abstract measure of inconvenience but an attempt to put a dollar figure on interrupted operations. A manufacturer losing production for several hours, a distribution center unable to move inventory, and a data center supporting continuous workloads do not carry the same outage exposure, and neither do facilities in regions with different reliability histories. The more useful comparison is increasingly total resilience cost against expected interruption loss, not battery cost against electricity savings alone.
Falling Battery Prices Widen the Resilience Case
The other side of that calculation is moving quickly, though the direction matters more than the headline number. BloombergNEF's $70/kWh figure reflects battery-pack pricing, not the fully installed cost of a storage system; companies still have to account for power conversion equipment, engineering, construction, controls, permitting and financing, and U.S. costs can run higher than global averages because of tariffs and supply-chain conditions. Still, BNEF's broader project-cost data shows the benchmark levelized cost of electricity for a four-hour battery storage project fell 27% in 2025 to $78 per megawatt-hour, its lowest level since the firm began tracking the technology in 2009, driven by lower pack prices, stronger manufacturer competition and improved system design.
The U.S. Department of Energy (DOE) is making a related argument about microgrids: facilities typically install them because they cannot tolerate significant outages, but the same systems can generate value during normal operation through time-of-use rate arbitrage, demand-charge reduction and optimized dispatch of onsite generation and storage. A resilience asset no longer has to sit idle waiting for an emergency to justify its cost, which changes how the investment should be modeled from the outset.
Finance and Facilities Need the Same Model
This is where the storage decision becomes less about technology and more about capital allocation. Facilities teams know which operations cannot tolerate an interruption, and finance teams know what those interruptions actually cost. Energy teams bring a third piece to the table: utility tariffs and the savings available on the grid-connected side. Those numbers need to be evaluated together rather than built into separate business cases that never meet, particularly as facilities also weigh the operational readiness steps utilities are already asking them to complete before peak season.
DOE's Federal Energy Management Program approaches resilience the same way through its Customer Damage Function Calculator, a free tool built to help facility owners establish the financial cost of a grid outage as the first step toward evaluating the avoided costs tied to resilience investment. For corporate buyers, that logic points to a more useful question than "what is the battery payback." The better question is what it costs a given facility when power disappears, how often that is likely to happen, and how much of that exposure a resilience investment can realistically avoid, a question that matters even more for organizations exposed to the longer restoration timelines that can follow when neighboring grid regions face simultaneous strain.
Storage will not pencil out everywhere. Batteries cannot protect every load indefinitely, and system design has to reflect outage duration, critical loads and available generation. But the hurdle has changed: with outage losses rising while stationary-storage costs fall, excluding reliability from the ROI calculation can make the cheaper investment on paper turn into the more expensive decision over the life of the facility.
