Research / Industry outlook
Energy infrastructure · Research outlook
Storage expands. Grid access sets the pace.
Project pipelines are growing, but connecting new storage to the grid is becoming the decisive constraint in this research sample.
In brief
- Fieldline’s project sample shows a growing storage pipeline: 8 GWh planned for 2024, 12 GWh for 2025 and 18 GWh for 2026.
- Developers identify the connection queue as the main schedule constraint.
- The project sample tracks planned capacity, not completed installations.
- Once a project has connection approval, transformer and other electrical equipment deliveries become the next constraint.
Storage developers are announcing more projects each year, and the projects are getting larger. That part of the story is familiar. What our sample shows more clearly is where those projects wait. For most of them, the wait is for a grid connection. Battery supply, the main worry two years ago, is now named by only a handful of developers as the thing holding a project back.
This outlook draws on Fieldline’s storage project sample: 163 utility-scale projects from 14 US developers, with target operation dates from 2024 to 2026. We surveyed the developers between August 17 and September 25, 2026, and asked them to name the main constraint on each project’s schedule. The sample is not a census of the market. It is a consistent panel that we can follow from one edition to the next.
Project pipeline
Fieldline’s fictional project sample contains 8 GWh of planned storage in 2024, 12 GWh in 2025, and 18 GWh in 2026. Developers in the sample identify the connection queue as the main schedule constraint. Announced capacity is a project pipeline, not completed installations.
Figure 1. Planned storage capacity in Fieldline’s project sample, by target operation year (GWh)
Planned capacity in the project sample. Not completed installations.
Source: Fieldline storage project sample, 14 developers, 163 projects, data as of September 30, 2026. Each project is counted once, in the target year stated when it entered the sample. The chart does not show completed installations.
| Target year | Projects | Planned GWh | Avg. MWh |
|---|---|---|---|
| 2024 | 41 | 8 | 195 |
| 2025 | 52 | 12 | 231 |
| 2026 | 70 | 18 | 257 |
| Total | 163 | 38 | 233 |
The pipeline grows by half at each step. Average project size rose from 195 MWh for projects planned for 2024 to 257 MWh for those planned for 2026, mostly because developers moved from two-hour to four-hour designs.
These figures describe plans. Each project is counted once, in the year of the target operation date it carried when it entered the sample, and later schedule changes do not move it. That rule keeps the series comparable from year to year, but it also means the bars cannot be read as capacity that came online. By September 30, 2026, developers had moved 3.1 GWh of the 2024 group and 4.4 GWh of the 2025 group to later target dates, and nine projects totaling 1.2 GWh had been withdrawn. We do not verify which projects reached commercial operation. The definitions, and what the sample leaves out, are in our methodology note on the project sample.
Grid connection
We asked developers to name the single main constraint on each project’s schedule. For 101 of the 163 projects, or 62%, the answer was the connection queue. No other constraint came close. Connection queues determine schedules in the developer sample.
Table 1. Main schedule constraint named by developers
| Constraint | Projects | Share |
|---|---|---|
| Connection queue (studies, upgrades, agreement) | 101 | 62% |
| Transformer or other electrical equipment delivery | 31 | 19% |
| Permitting and land use | 17 | 10% |
| Battery supply | 9 | 6% |
| Other (financing, offtake) | 5 | 3% |
Source: Fieldline developer survey, August 17 to September 25, 2026. One answer per project, given by the developer.
The queue is slow for structural reasons. A request is studied together with the others filed in the same window, and the studies often find that the network needs upgrades before new capacity can connect. Each round of results can change a project’s costs, and projects that drop out force restudies for the ones that remain. Our explainer on how connection queues work walks through the steps.
The waits are getting longer. For projects in the sample that have signed an interconnection agreement, the median time from request to signed agreement was 25 months for projects planned for 2024 and 33 months for projects planned for 2026. Operator data point the same way. Midland Grid’s third-quarter queue update reports an average wait of 34 months for standalone storage agreements signed in the year to September 30, up from 27 months two years earlier. Storage now accounts for about a third of the capacity in that queue.
Several developers told us they now plan schedules backward from the expected agreement date rather than forward from site control. One called the battery order “the easy part of the schedule now.” The queue also changes how projects are announced. Developers file earlier, often before the first study results arrive, because a queue position has value of its own. That is one reason announced capacity grows faster than capacity that reaches operation.
Supply constraints
Once a project has its agreement, the constraint moves. Of the 58 projects in the sample that already hold a signed interconnection agreement, 24 named transformer or other electrical equipment delivery as the main constraint on their schedule. For those projects the question is no longer when the grid will accept them but when the equipment to connect them will arrive.
Supplier data agree with what developers told us. Vale Systems’ storage equipment supply outlook, published October 6, reports transformer lead times that reach 18 months for the projects in its sample, against a median of five months for battery enclosures. Main power transformers for project substations are the longest item, followed by high-voltage breakers and pad-mounted step-up transformers.
Battery supply has moved the other way. Cell prices in Vale’s procurement data fell 19% in the year to the third quarter, and developers in our sample described battery orders as routine. Seven of the nine projects that named battery supply as their main constraint were planned for 2024. Lower prices improve project economics, but they do not bring a connection date forward.
Two cautions apply when reading the supplier and developer data together. First, they measure different things. Vale’s sample covers equipment delivery for projects that have reached the ordering stage, while ours covers the whole pipeline, including projects that are years from an agreement. Second, a long lead time delays a project only if the order is placed late. Developers who ordered transformers early, in some cases before the agreement was signed, reported fewer equipment delays. They accepted cancellation costs if the project fails in study.
Our reading is that the queue will remain the main constraint for most of the sample through 2026, and that equipment delivery will matter more for the growing share of projects that clear it. In the next edition we plan to add a check on operating status for the 2024 group.