Edition 10 · 26 August 2026 · 5 min read
The AER published its wholesale market performance review last Thursday. Prices fell in every region and in every time-of-day block last year, which is the line most of the coverage ran with, and it's close to the least useful sentence in the document. The number worth your time is where the money went.
The Take: eighty-two per cent of the money arrives after 4pm
Total NEM energy revenue fell from $23.8 billion in 2024 to $19.6 billion in 2025, down 18%. That drop is real and it's welcome. But 82% of what remains was earned in two windows: 49% in the evening peak between 4pm and 9pm, and 33% overnight. The middle of the day, where prices fell hardest, is where almost none of the money is.
Run the series back further and it sharpens. Evening peak revenue was $6.4 billion in 2021. In 2025 it was $9.6 billion, up 50%, inside a market whose total revenue is still 43% above its 2021 level of $13.7 billion. Average prices came down last year. The concentration didn't.
The AER's own framing is that the NEM is transforming from one market into many within each region, split by time of day and by service type. That's a regulator being unusually direct about structure, and it explains why the flat rate still sitting under most commercial contracts has quietly stopped describing anything. A single c/kWh figure is an average of blocks that are now moving in opposite directions.
Who sets that price has changed too. Battery generation and battery charging together set the wholesale price in 16.3% of dispatch intervals across the NEM in 2025, up from roughly 1% in 2021, displacing gas and hydro. In Queensland's evening peak it was 25.5%. The marginal unit in the most expensive hours is now an arbitrage decision rather than a fuel cost. If you've been reading gas forwards as a proxy for where peak prices go, I'd stop.
Here's what I'd change before the next renewal. Stop treating a flat rate as the unit of comparison. Hand your retailer your actual interval data and ask them to price the evening block, the overnight block and the daytime block separately, then compare offers block by block instead of on the blend. If your load is genuinely daytime-weighted, you're sitting on the cheapest part of a market that has repriced around you, and a flat quote hands that discount straight back to whoever wrote it. If your load is evening-weighted, you want that priced where you can see it, because the evening peak is exactly where the AER says competition risk has relocated, and a number on the page is a number you can argue with.
Quick Hits
Texas paused new data centre development and the EIA cut its load forecast by more than half. In the 11 August Short-Term Energy Outlook, the EIA revised Texas electricity load growth for 2027 down from 14% to 6%, attributing the change directly to the Texas governor's 3 August announcement of a pause on new data centre development. So what: the biggest single assumption in global demand forecasting lost eight percentage points in one state on one political decision, which is a useful reminder that data-centre load is a planning-permission variable before it's a physical one.
A $1.1 billion cut to network revenues is sitting in a December decision. The AER's draft 2026 Rate of Return Instrument, published 29 May, estimates that its proposed 0.37 percentage point reduction to the equity risk premium would cut regulated network revenues by roughly $1.1 billion in aggregate. Submissions closed on 31 July and the final instrument is due in December. So what: network charges are the largest non-energy component of a commercial bill, and this instrument sets the allowed return on every regulated network's asset base for the next round of determinations. If you're modelling network cost past 2027, that's the date to hold.
The threshold for large inverter-based load connections could go from 5 MW to 30 MW. A draft AEMC rule on technical access standards proposes lifting the distribution-level threshold six-fold, with a final determination due late October. So what: if you're planning electrification, on-site storage or any material load increase at a distribution-connected site, the standard you'll have to meet is being rewritten right now and the consultation is still open.
From the blog
If the Take has you reopening a contract, our guide to energy procurement for large business covers the structures underneath it: how the main contract types differ, how to time a tender against a falling forward curve, and where Scope 2 weighting changes the evaluation. Read it here.
The Sign-Off
AEMO's Electricity Statement of Opportunities for 2026 hasn't landed yet. When it does I'll be reading it alongside this report rather than on its own, because a reliability forecast tends to smooth over exactly the thing the AER has just documented: a system where the value has collected into a handful of hours. Same grid, two very different pictures of where the pressure sits, and the gap between them is where most procurement mistakes get made.
If you've priced a contract by time block this year and the answer came out somewhere you didn't expect, tell me. I want to know whether the retail market has caught up with this yet.
Until next fortnight,
Cohen
Find me on LinkedIn.
Get Behind the Meter in your inbox every fortnight. Subscribe →