AEMO checked home batteries again, across about half a million homes: 74 per cent less grid power at the evening peak

One quarter of measured data is a finding. Two quarters, the second from a sample about fifty times larger, starts to look like a fact about how home batteries behave. That is what AEMO’s September-quarter report adds, and it is the part of the report that bears most directly on a household thinking about a battery.

The same answer from a much bigger sample

The first comparison was small and deliberately narrow: detached houses with under 20 kW of solar and, for the battery group, only batteries installed between 1 July and 1 December 2025. This quarter AEMO drew on detached houses with solar in its register of distributed energy resources, about 2.4 million metering points in all, with the battery group taking in every system installed before 1 July 2026. AEMO’s summary is that households with rooftop solar and batteries drew around three quarters less power from the grid during the evening peak than households with solar alone.

The daily shape matched the first reading too. Battery homes went into net export later in the morning, because the battery fills before any surplus goes to the grid, and through the evening peak they stayed net exporters for longer and imported less.

Our view. A result that holds when the sample grows about fifty-fold, and when the battery group widens from five months of installs to everything installed before July, is the strongest sign so far that the evening effect is real rather than a quirk of which homes were picked. Two cautions remain. It is a comparison, not a trial: households that bought a battery may differ from those that did not in ways that also change their evening use, and the report describes no adjustment for that. And the two quarters are different seasons, autumn into winter and then winter into spring, so two matching readings are encouraging rather than settled. The figure measures grid power avoided, not money saved, which depends on your tariff.

At grid scale, batteries now set the price almost half the time

The big batteries tell the same story at a different scale. Energy shifted by grid-scale batteries more than tripled on a year earlier: on an average day they charged 13.4 GWh between 10am and 4pm and discharged 10.9 GWh between 4pm and 9pm, against 3.7 GWh and 3.2 GWh in the same quarter of 2025. Through the evening peak they supplied 8.0 per cent of all generation, more than gas at 5.1 per cent, which AEMO says is the first time that has happened in any quarter.

Apr to Jun 202517% Apr to Jun 202636% Jul to Sep 202518% Jul to Sep 202648%

Battery charging and discharging combined, NEM-wide. Each pair is a quarter and the same quarter a year earlier; October 2025 to March 2026 is not shown. Sources: AEMO, QED Q2 2026 and QED Q3 2026.

Fewer free-power hours in the middle of the day

Negative or zero wholesale prices occurred in 14.4 per cent of dispatch intervals, down 4.5 percentage points on the same quarter last year and lower in every mainland region. New South Wales fell 9.1 points to 5.2 per cent and Queensland 12.1 points to 13.8 per cent. Tasmania went the other way, up 9.2 points to 13 per cent. Part of AEMO’s explanation is the batteries again: renewable generators usually set negative prices, while battery charge was more likely to set prices at positive levels. The trend was already visible last quarter, when negative prices got rarer and shallower.

Our view. Most retail plans charge a set rate per kilowatt-hour whatever the wholesale price is doing, so for most households this is invisible. If you are on a plan that passes wholesale prices through, or you have timed loads to chase negative midday prices, those hours are getting scarcer, and the grid’s own operator now names battery charging as one reason.

One record the batteries did not set

Minimum demand on the grid fell to a new all-time low of 9,422 MW across the market on Sunday 20 September, and Queensland set its own all-time low of 2,741 MW the same day. AEMO’s explanation is rooftop solar and the weather: Higher distributed PV output and mild spring conditions led to new minimum operational demand records. Rooftop solar output averaged a record 3,203 MW for a September quarter, up 13 per cent. That is a solar story, not a battery one.

Wholesale prices fell again. Your bill moves later, if at all.

Wholesale prices across the market averaged $63/MWh, down 27 per cent on a year earlier and 15 per cent on the June quarter, the lowest September-quarter average since 2021. They ranged from $56/MWh in Victoria and Tasmania to $74/MWh in New South Wales; Queensland averaged $61/MWh and South Australia $69/MWh. AEMO puts the fall down to lower evening peak demand, more battery participation and less price volatility.

As we said last quarter, wholesale energy is one part of a retail bill alongside network charges, environmental schemes and retail costs, and default prices are reset once a year. Two cheaper quarters in a row is a better sign than one, and it shows up in a bill late or not at all.

Home batteries keep going in

Citing the Clean Energy Regulator, AEMO reports cumulative home battery capacity under the Cheaper Home Batteries Program in the market’s five regions of 13,917 MWh at the end of September, up 1,652 MWh or 13 per cent in the quarter, with more than 10 per cent growth in every region. New South Wales had 5,885 MWh, Victoria 3,185 MWh, Queensland 3,166 MWh, South Australia 1,568 MWh and Tasmania 114 MWh. AEMO counts by market region from the regulator’s postcode data, which includes only installations whose certificates have been approved, so its totals do not match the regulator’s own state figures in our 21 October story.

Sources

Where this came from. Researched, written and kept up to date by AI systems run by a small Australian team, and built from the primary sources below. How The Electric Household works.

1. AEMO, Quarterly Energy Dynamics Q3 2026 (PDF, 90 pages, published October 2026, downloaded and read 28 October 2026): the household solar and battery comparison, the 0.7 kW and 74 per cent figures and the sample in footnote 8, grid-scale battery energy shifting, evening peak share and price-setting share, negative price occurrence and AEMO’s explanation of it, minimum demand records, rooftop solar output, wholesale prices by region, and home battery capacity by region.
2. AEMO, Quarterly Energy Dynamics: the report series landing page, where the Q2 2026 report is also published (the 17 and 36 per cent price-setting figures and the 10,000-household samples).
3. The Electric Household, AEMO measured what a home battery does to your evening peak: the first reading, from the Q2 2026 report.
4. The Electric Household, South Australia’s grid demand fell to 57 MW on a Saturday afternoon: negative prices in the June quarter.
5. The Electric Household, Since the 1 May rebate change, most rebated home batteries are 14 to 28 kWh: the regulator’s own installation figures.

Methodology. Every figure is from AEMO’s Quarterly Energy Dynamics Q3 2026, covering 1 July to 30 September 2026; we downloaded the 90-page PDF on 28 October 2026 and read it rather than a summary of it. The comparison figures for the June quarter (the 10,000-household samples, the 73 per cent reduction and the 17 and 36 per cent price-setting shares) are from AEMO’s Q2 2026 report, which we read for our 3 August story. The household comparison is AEMO’s own, from National Metering Identifier sample datasets: detached dwellings with solar recorded in AEMO’s DER Register, with a battery group of systems installed before 1 July 2026. The “about fifty times larger” is our arithmetic, 0.5 million against 10,000 battery households, and AEMO gives its sample sizes only approximately. We have not checked the sampling independently. The evening peak is AEMO’s definition of 1600 to 2100 hours. The 0.7 kW figure is a quarterly average, not a promise about any one home. The battery chart pairs the only four quarters whose figures we have read; it is not a continuous series. The paragraphs labelled as our view are ours, not AEMO’s. We did not ask AEMO for comment.

Have a battery and read your own data differently? Tell us and we will check it against the report and log the outcome here.