Data

The community battery rollout planned 92 by the end of last year and has built 22. Some are working the voltage backwards

Community batteries are the answer offered to households that cannot put storage on their own roof: renters, apartments, shaded blocks. ARENA has now published the first portfolio-wide data report on its programme, prepared by enX and covering the second half of 2025. Across the portfolio, 284 batteries are planned and 23 are installed. Within the slippage analysis, initial planning projected 92 installed by 31 December 2025; the actual figure is 22. Capital expenditure is now expected to reach $480 million. And the interval data shows some batteries charging when network voltage is low and discharging when it is high, which is the opposite of helping.

30 August 2026 Report read in full 30 Aug 2026

What the report is, and how much weight it carries

The ARENA Community Battery Portfolio Data Report 1 was prepared by enX, is dated 25 March 2026 and covers the reporting period 30 June to 31 December 2025. It draws on survey responses from 13 of 16 proponents, covering 284 community batteries, plus operational interval data for 28 batteries supplied by three proponents: Shell Energy, Ausgrid and Hydro Tasmania. Shell Energy's interval data was limited and was excluded from this analysis.

Its own framing is modest. It exists to give a point-in-time overview of the current and planned community battery portfolio and to inform a discussion about where further analysis should go. And it states the caveat that governs everything in it: All figures and findings presented are based solely on information supplied by proponents and should be interpreted in this context. These are self-reported numbers from the organisations building the batteries.

Two installed counts, and they are not a contradiction

The report gives 23 batteries installed in one place and 22 in another, and the difference is scope rather than error. 23 is the portfolio-wide figure, against 227 batteries are expected to be installed by the end of 2026 and 284 installed by 2027. 22 appears inside the project-slippage analysis, which covers only the 9 of 15 proponents that have submitted survey data in at least two reporting periods. The 92 figure comes from that same subset, so 92 against 22 is a like-for-like comparison within it.

The community battery portfolio, from ARENA Community Battery Portfolio Data Report 1
Measure	Figure
Batteries covered by survey data	284, from 13 of 16 proponents
Installed, portfolio-wide	23
Planned by end of 2026	227
Planned by 2027	284
Slippage subset: planned by 31 Dec 2025	92
Slippage subset: actually installed	22
Expected portfolio capital expenditure	$480m, up from $419m
Technology and EPC share of capex	87%
Batteries with operational interval data	28

Who slipped, and who did not

The report names them. The proponents that have experienced delays in completing their first install are SA Power Networks, Endeavour Energy, Horizon Power, Tasmanian Networks, Western Power and Transport for NSW. Energy Queensland is the only proponent that has brought forward their planned first install, expected by 1 July 2026.

The forward number has nonetheless gone up rather than down, and that is worth stating alongside the slippage. The total expected by the end of 2027 has increased, the report says, primarily due to Ausgrid's effort to increase their planned battery installs from 16 to 46. Ausgrid is expected to account for 81.7 per cent of the portfolio's total capacity by 2028, or 538 MWh. One network is carrying most of this programme.

The money

Expected capital expenditure across the portfolio is now $480 million, up from $419 million at the previous reporting period. The report attributes that rise not to cost blowouts but to larger survey response rates and proponents firming up the budget of their project, which is a more benign explanation and one we have no basis to doubt. Technology and EPC costs account for 87 per cent of portfolio-wide capital expenditure, so this is overwhelmingly a hardware-and-installation programme rather than a soft-cost one.

The operational finding, stated carefully

This is the part that has not been reported anywhere else, and it needs its qualifiers kept attached.

The broad pattern is the right one. Across the smaller batteries, those rated under 4,000 kW, the report finds they are often discharging during the morning and evening peaks and are charging during the middle of the day, which is exactly what a battery in a solar-heavy network should do: soak up the midday surplus, give it back at the peak. The larger batteries show the same shape with less morning and overnight activity and a sharper discharge profile during the afternoon peaks.

Then the qualifier. Some batteries are charging when voltage is low and discharging when voltage is high. This indicates some batteries are struggling with voltage management. The report immediately balances it: However, other batteries have a flatter voltage profile and are demonstrating good voltage management.

Why it matters for a household rather than an engineer: high network voltage in the middle of the day is the thing that trips rooftop solar inverters off and costs a household its export. A battery charging then pulls voltage down and keeps the street's solar working. A battery doing the reverse is not neutral, it is working against the problem it was funded to help with. The report does not say how many are in each group, and we are not going to guess a proportion from a chart described in words.

The engagement finding nobody will quote

Buried in the survey results is a count of how proponents engage the communities these batteries are named after. A project webpage is the most common approach, at 12 responses. Newsletters follow at 7, door-knocking or community stalls and public meetings at 6 each, walking tours at 5, an advisory or reference group at 4, and social media pages and co-design workshops at 3 each, which the report notes are among the least frequent.

Take that for what it is, a count of approaches rather than of effort or quality. But the pattern is that the most common form of community engagement in the community battery programme is publishing a webpage, and the least common is designing the thing with the community.

Our view

The deployment gap is the headline and it should be read without drama. Infrastructure programmes slip, network connection processes are slow, and the same report shows the forward pipeline growing rather than shrinking. Nobody has abandoned this. But 22 against a planned 92 inside the subset that has reported twice is a large enough gap that a household waiting for a battery on their street should plan on years rather than months, and six named network businesses missing their first install is a pattern rather than one unlucky project.

The finding we would actually act on is the voltage one, because it is the closest thing here to a test of whether the concept works. A community battery earns its subsidy by making the local network better for the houses around it. The report's own data says some are and some are not, from a sample of 28 with one proponent excluded. That is a small and preliminary finding and it is also the first portfolio-wide operational evidence there is, which is why it deserves the next report rather than a press release.

What we would want next: how many of the 28 fall into each voltage group, whether the six delayed proponents have a common cause, and whether the $480 million holds once the 227 due by the end of 2026 are actually built.

Sources

enX for the Australian Renewable Energy Agency, ARENA Community Battery Portfolio Data Report 1 (PDF, dated 25 March 2026, downloaded and read in full 30 August 2026): the reporting period of 30 June to 31 December 2025; survey responses from 13 of 16 proponents covering 284 community batteries and operational interval data for 28 batteries from Shell Energy, Ausgrid and Hydro Tasmania, with the quoted exclusion of Shell Energy's data; the quoted purpose of the report and the quoted caveat that all figures are based solely on information supplied by proponents; the quoted statement that 227 batteries are expected by the end of 2026 and 284 by 2027 and that 23 have been installed; the project slippage analysis covering 9 of 15 proponents, the initial planning figure of 92 installed by 31 December 2025 against 22 actual, the named proponents that have experienced delays, and the quoted statement that Energy Queensland is the only proponent to have brought forward its first install; the quoted attribution of the increased forward total to Ausgrid raising planned installs from 16 to 46, and that Ausgrid is expected to account for 81.7 per cent of portfolio capacity by 2028 at 538 MWh; the $480 million expected capital expenditure, its rise from $419 million and the quoted explanation for that rise, and the 87 per cent technology and EPC share; the quoted findings on battery power and voltage for batteries rated under and over 4,000 kW, including the quoted statements on batteries struggling with voltage management and on others demonstrating good voltage management; and the community engagement response counts.
ARENA, Knowledge Bank entry for the report (read 30 August 2026): the 18 June 2026 publication date and ARENA's own summary of the report's scope.

Methodology. The report was downloaded from ARENA's Knowledge Bank and read in full; every figure and quotation is from it. Where the report gives two installed counts we have explained the scope difference rather than picking one: 23 is portfolio-wide and 22 sits inside the slippage analysis covering 9 of 15 proponents, which is also the source of the 92 planned figure, so that comparison is internally consistent. ARENA's own summary of the report describes 23 operational batteries; we have used the report's own figures throughout. The voltage findings come from interval data for 28 batteries with one proponent's data excluded, and the report does not state how many batteries fall into each group, so neither do we. All figures are self-reported by proponents, as the report states. We have not contacted ARENA, enX or any proponent.

The report carries a disclaimer from enX that it is provided without guarantee, representation, condition or warranty of any kind, and that any party relying on it assumes the entire risk as to its accuracy, currency or completeness.

Work on one of these projects, or read the interval data differently? Request a correction and we will check it against the report and log the outcome here.