Long-duration storage: what the evidence shows about its role in the grid

Long-duration storage is often presented as a technology category, but the term alone does not describe one uniform solution. Operationally, it refers to systems able to deliver electricity for longer periods than short-duration storage applications. The threshold, however, depends on the programme, the problem being addressed and the technology being assessed. Comparing projects without stating how many hours count can therefore lead to misleading conclusions.

A Horizon Europe call focused on demonstrating long-duration batteries defines its scope as battery storage concepts lasting more than eight hours at system level. This figure describes the scope of that call; it does not establish a universal definition for every policy, technology or market. Nor does it mean that eight hours is the technical boundary between categories, or that a system exceeding the threshold has proved its usefulness on a real grid. Source

The essential distinction is between power and energy. Power indicates how much a system can deliver at a given moment; the available energy, combined with that power, determines how long delivery can be sustained. A project may have high power and only a few hours of storage, or lower power and longer duration. Published duration should be read alongside operating conditions: discharge power, initial state of charge, usable energy and the criterion used to declare that discharge has ended.

Official funding documents help identify which problems governments aim to address and what kinds of demonstrations they consider relevant. The United Kingdom, for example, publishes projects selected in one phase of its longer-duration energy storage demonstration programme, grouping initiatives into categories that include thermal storage. The existence of a project page or its selection for funding confirms an institutional process; it is not equivalent to independent validation of final performance. Source

The European call for demonstrations of batteries lasting more than eight hours focuses on assessment at system level. That framing matters: the performance of an isolated battery does not answer every question about operation, integration, controls, safety, maintenance and response to grid needs. But a research priority or open call does not prove that funded projects have already delivered field results either. An announced test is evidence of intent and design; the test result requires published data. Source

Rigorous reading separates three levels that are often blurred in promotional material: a proposed technology, a facility that has been built or funded, and a system operated with documented results. Moving from one level to the next requires new evidence. A projected figure for a future plant should not be presented as measured performance; a pilot demonstration does not automatically show that a system is competitive at commercial scale or can be replicated on another grid.

Long-duration storage covers different mechanisms. Batteries store energy through electrochemical processes; thermal solutions retain heat or cold for later use, while other designs may convert electricity into potential energy, pressure or fuels and recover it later. The UK programme source explicitly identifies thermal-storage projects, while the cited European call focuses on battery systems. These examples show that funding categories are not interchangeable and, by themselves, do not provide a performance comparison between technologies. Sources

Round-trip efficiency—the share of energy put in that can be recovered as electricity—can help compare electrical systems, but the measurement boundary must be specified. If the calculation excludes auxiliary consumption, inverter losses, waste heat or energy used by peripheral equipment, two values are not directly comparable. For thermal storage, the service may also be the delivery of useful heat rather than electricity: measuring it only by conversion back to electricity might not represent its intended use. Such differences should be made explicit, not hidden behind a single figure.

Nominal power must also be distinguished from sustained power, nominal energy from usable energy, and theoretical discharge hours from the hours actually available in a specific operation. A useful assessment states whether duration was calculated at constant power, what charge reserve was held back, how many cycles were completed and whether the system met grid-connection constraints. Without comparable conditions, the number of hours is a descriptor, not a ranking of superiority.

A responsible comparison does not rely on a single figure. It should connect the service to be provided with system characteristics and the economic context. At a minimum, look for information about discharge duration and power, efficiency measured with stated boundaries, service life or degradation under an identifiable protocol, relevant total cost, and site and interconnection requirements. Missing one of these data points does not make a project useless, but it limits the strength of the conclusions that can be drawn.

A cost figure per unit of capacity does not necessarily answer how much it costs to deliver electricity during the hours and under the conditions a grid needs. The scope and calculation basis matter: equipment cost or total system cost, upfront investment or lifetime cost, and assumptions about financing, maintenance, replacement, charging and utilisation. The cited innovation calls provide information about programmes and projects, but the pages cited do not constitute a consistent cost database and do not allow technologies to be ranked by price or profitability.

The following list is an initial screen for project reports and announcements. It is not a formal standard and does not replace analysis of the use case:

• Duration and power: stated hours, sustained power and usable energy, with the calculation method.
• Efficiency: input and output energy, included losses and test conditions.
• Durability: cycle count or operating time, degradation and warranty conditions, if published.
• Economics: what the cost figure covers and the assumptions about utilisation, maintenance and replacement.
• Integration: space, connection, infrastructure, safety and the grid service the project intends to provide.

Longer duration may matter when the problem requires shifting energy between times several hours apart or sustaining delivery for longer. But duration alone does not show whether an installation will solve a specific supply problem: the time it charges, available energy, the grid’s ability to transport that energy and the timing of demand all matter. The programme sources reviewed document demonstration objectives, not a general conclusion about which technology will meet the needs of each electricity system.

The scale of analysis also changes the answer. A system may be technically capable of discharging for many hours, yet useful operation depends on factors such as the availability of charging energy, siting, permits and interconnection. Documentation of a demonstration should not be treated as proof of broad commercial deployment, and a description of a selected project does not establish continuous operation or its net contribution to the grid if those results have not been published.

Consequently, storage should be assessed against a specific need and alongside other available options, not by an isolated technology label. This analysis does not support the claim that a particular duration is sufficient, or that one technology family is universally preferable. Usefulness to the grid is a property of the system and its operating context, not just of the storage chemistry or mechanism.

The official pages reviewed provide clear references for two points: the more-than-eight-hour threshold used by a European battery demonstration call, and the existence of UK projects selected for demonstration, including thermal projects. They do not provide a consistent comparison of efficiency, cost, service life and commercial operation across different technologies. Nor, on their own, do they support a conclusion about which solution will be cheaper, more reliable or better suited to a particular grid. These limits do not invalidate the programmes; they define what their documents can demonstrate.

To move from objectives to performance evidence, a report should identify the system and its scale; the test period and conditions; the power and energy actually delivered; efficiency and auxiliary consumption; operating and degradation data; and the calculated cost, with its scope and assumptions. For a connected installation, it is also useful to distinguish site-measured results from simulations, projections or manufacturer data. When these elements are missing, conclusions should remain at the level of a promise, design or announced demonstration.

The practical question is not simply whether a technology offers many hours, but what service it delivers, under what conditions and at what verifiable total cost. Until sources publish such data in a comparable form, editorial caution means describing objectives and tests precisely, attributing figures to their context and avoiding the equation of funding selection with technical or economic success.