Data

Australia ran a two-year trial of 51 all-electric homes. Year two contradicted year one

Almost everything you read about all-electric housing is a projection. This is not. From October 2020, Frasers Property and ARENA monitored 51 Net Zero Energy Demand homes at Ed.Square in Edmondson Park against 10 business-as-usual dwellings, and the final memorandum is now public. The emissions finding held across both years. The electricity-use finding reversed: in year two the all-electric homes used more, and the conventional ones slightly less. The most useful thing in the document is how plainly it says what it cannot tell you.

30 August 2026 Memorandum read 29 Aug 2026

What was actually built and measured

In October 2020 Frasers Property Australia, with ARENA, launched a $1.42 million project to deliver 51 Net Zero Energy Demand homes at Ed.Square, its masterplanned community at Edmondson Park in south-western Sydney. The point was not to build them and declare victory. It was to monitor their energy and environmental performance and compare the results against 10 conventional dwellings on the same estate. The project partner was Real Utilities, and the analysis was done by Arup.

The final memorandum, dated 30 April 2026 and covering March 2025 to March 2026, confirms the findings of the earlier Milestone 3 report and adds a second year of data. Two years of overlapping operational measurement of real occupied houses is rare, and it is the reason this document is worth more than any modelled comparison.

What held, and what reversed

The report gives three key findings after combining both years. The first is unambiguous: All-electric homes are better for immediate and long term operational emissions. That is the central claim for electrification and the trial supports it.

The second is a caution that gets less airtime: All-electric homes may present a higher grid peak demand depending on systems used, however local PV, batteries and home energy management systems could reduce this. Peak demand and total consumption are different problems, and the trial says the first can get worse even where the second does not.

The third finding is the one that changed. In year one the all-electric homes showed lower average demand than the conventional ones. In year two the available data showed an increase in electricity consumption across the NZED homes, and it was primarily driven by hot water systems, plug loads and a small number of high consuming households, while the business-as-usual homes showed a slight reduction in overall electricity use. The report states the obvious consequence itself: This represents a departure from year 1 trends, where NZED homes showed lower average demand.

Note what is doing the work in that sentence. Hot water and plug loads are not the building. They are the appliances and the people. A small number of high-consuming households can move an average across 51 homes, and the report says so rather than smoothing it away.

The occupant survey, where the reversal is sharpest

A follow-up survey drew 36 responses, of which 4 were excluded because the properties were rented, leaving 32 valid responses against 33 in the first survey. On comfort, approximately 53 per cent of all-electric respondents found their cooling system effective at maintaining a comfortable temperature in summer, and on average about 52 per cent found their home comfortable with no heating or cooling at all.

On bills, the numbers moved a long way.

Satisfaction with the affordability of monthly energy bills, Ed.Square occupant surveys
Survey	All-electric (NZED)	Business as usual
Survey 1	52%	12.5%
Survey 2	53%	57%

The all-electric figure barely moved. The conventional figure went from 12.5 per cent to 57 per cent, and the report calls it a notable shift. That is a four-and-a-half-fold change in one group across roughly a year, on a base of about 30 responses split between two groups.

We would not read that as the conventional homes becoming cheaper to run than the electric ones. A satisfaction question measures how people feel about a bill against their expectations, and between the two surveys the whole country's energy prices, rebates and news coverage moved. On a sample this size a handful of households changing their answer produces exactly this swing. What it does show is that occupant sentiment is not a proxy for building performance, and anyone quoting the first survey's 52-against-12.5 as evidence should know the same question returned 53-against-57 a year later.

The limitations, which the report states better than most

It would be easy to run this trial as a headline in either direction. The reason not to is printed in the document. On the sample, the report says a normalised methodology was applied as a best-effort approach, but that the results should not be considered a statistically robust representation of system efficiency, because of the limited and varied sample and the challenges associated with data completeness and quality.

Those challenges are itemised rather than gestured at. The project hit Weak signal and connectivity issues, System outages, and Accidental customer interaction with monitoring equipment, which together resulted in intermittent data availability, reducing the reliability of comparisons both within and between the housing typologies.

Year two specifically had holes in it. No new PV data was available, year 1 values were applied on an aggregate basis, and gas metering for the conventional homes was incomplete, with the data ending in September 2025, limiting full year comparisons. So the year in which all-electric consumption rose is also the year with reused solar figures and a truncated gas series on the comparison group.

On the heating and cooling technologies, the report tested variable refrigerant flow against ground source heat pumps and declined to call it: the VRF systems demonstrated more efficiency, and lower peak loads, than GSHP systems in this study, however the sample size was too small, and there were confounding variables, preventing robust conclusions. A finding withheld is worth more than one asserted, and this is the second time in this document the authors have refused to say something their own data hinted at.

What we take from it

Our reading is that this trial is more valuable for what it settles than for what it reverses. The emissions case for all-electric housing survived two years of real measurement, and that was the claim most worth testing. The consumption comparison did not survive, and the honest conclusion is not that all-electric homes use more power. It is that a 51-against-10 comparison with connectivity dropouts and reused PV data cannot answer the question either way, and the people who ran it say so.

The finding we would carry into your own decision is the report's own conclusion, which is about neither emissions nor bills: technology alone does not guarantee performance, with Early-stage planning, integrated design, selection of systems and occupant behaviour named as the things that decide the outcome. Hot water and plug loads moved the year-two average in a purpose-built estate of net zero energy demand houses. In an ordinary house retrofitted one appliance at a time, they will move it more.

The other practical takeaway is the peak demand caution. If you are going all-electric and your hot water and car both draw in the early evening, you may be cheaper to run and worse for the grid at once, which is what timers, batteries and a home energy management system exist to fix. The trial found that too, and it is the least reported of its three findings.

What we would want next and do not have: the underlying monitoring dataset rather than the memorandum, and whether the year-two rise persists once the PV data gap is filled. The memorandum is a summary of an analysis, not the analysis.

Sources

Frasers Property Australia and ARENA, Net Zero Energy Demand Homes: Milestone 4 Memorandum (PDF, 11 pages, dated 30 April 2026, downloaded and read in full 29 August 2026): the October 2020 launch, the $1.42 million project value, the 51 NZED homes and 10 business-as-usual comparison dwellings at Ed.Square, Edmondson Park; the March 2025 to March 2026 reporting period and contract 2019/ARP045 under ARENA's Advancing Renewables Program; Real Utilities as project partner and Arup as the analyst; the three quoted key findings on emissions, peak demand and the VRF against GSHP comparison; the quoted year-two observations on increased NZED consumption, its drivers, and the reduction in BAU electricity use; the quoted departure from year 1 trends; the survey response counts and the comfort and bill-affordability percentages including the quoted notable shift; every quoted limitation, being the statistical robustness caveat, the three data collection challenges, the intermittent data availability, the missing year 2 PV data and the gas metering ending September 2025; and the quoted conclusion on technology alone and the four factors that decide performance.
ARENA, Knowledge Bank entry: Frasers Property, Net Zero Energy Demand Homes, Final Memorandum (read 29 August 2026): the 16 July 2026 publication date and ARENA's own one-line description of the report's scope.

Methodology. The memorandum was downloaded as a PDF from ARENA's Knowledge Bank and read in full; every figure and quotation here is from that document. Percentages are the report's own. The one piece of arithmetic that is ours is the description of the business-as-usual bill-satisfaction change as roughly four-and-a-half-fold, which is 57 divided by 12.5. Where we say the survey swing should not be read as a cost comparison, that is our judgement and it is labelled as such, not a finding of the report. We have not contacted Frasers Property, ARENA or Arup, and we have not seen the underlying monitoring dataset.

The report carries its own disclaimers, including that Frasers Property makes no warranty as to the accuracy or completeness of the information in it, and that the views expressed are not necessarily those of the Australian Government.

Live in one of these homes, or work on residential monitoring? Request a correction and we will check it against the memorandum and log the outcome here.