Edition 8 · 28 July 2026 · 5 min read
Almost every commercial energy strategy I've looked at in the last five years rests on the same assumption: the money is in the evening peak. AEMO's Q2 2026 numbers, out last Friday, say that quietly stopped being true about a year ago. Batteries got there first.
The Take: the evening peak stopped paying
Grid-scale batteries set the NEM spot price in 36% of dispatch intervals last quarter, up from 17% in Q2 2025 (AEMO Quarterly Energy Dynamics). Charging and discharging combined, that makes batteries the most frequent price-setting technology in the market, ahead of coal, gas and hydro. It's a structural change in who decides what you pay, and it happened in about eighteen months.
Here's the number that actually lands in a budget. AEMO splits the average spot price into an energy component, which is the spot price capped at $300/MWh, and a cap return component, which is everything above that. The NEM cap return fell to $3/MWh last quarter, down 92% from $37/MWh a year earlier. In New South Wales, Queensland and Victoria it was effectively zero. South Australia was the only mainland region with a meaningful cap return, and nearly all of it came from two days, 21 and 22 June, which on their own contributed $13/MWh to SA's quarterly average.
The mechanism isn't mysterious. Battery capacity grew by 4,640 MW over the year, more than doubling to over 9,000 MW. Daytime charging rose 211% and evening discharge rose 228%. Average discharge hit a record 476 MW, almost triple the year before. The result, in AEMO's words, is a significantly flatter intraday price profile: daytime prices broadly unchanged, evening prices well below last winter. NEM average spot landed at $74/MWh, down 47% year on year and the lowest Q2 since 2020.
I'd be careful about declaring the evening peak dead. Those two days in South Australia are the reminder. Batteries have arbitraged away the ordinary expensive evening. The genuinely scarce one is still out there, just rarer and lumpier than it used to be. The risk changed shape rather than disappearing.
So what would I do with this before the next renewal? Two things. Re-run your interval data against this year's intraday shape rather than the historic one, and re-price the peak component specifically. If you hold cap products or are being sold them, check what the cap return actually delivered across your regions over FY26, because in three of the five it delivered close to nothing. Then look at timing. ASX FY27 base futures averaged $81/MWh across the mainland last quarter, down 16% year on year, but FY28 and FY29 fell by less, leaving every region's forward curve sloping upward. The discount sits in the near year. I'd take FY27 shape now and stay short on the out-years rather than reflexively going long into a falling market.
Quick Hits
A third of Britain's meters are now settled half-hourly. More than 11.3 million smart meters had migrated to Market-wide Half-Hourly Settlement by mid-June, roughly a third of Britain's metering systems, with full migration still targeted for May 2027 (Clean Flexibility Roadmap). So what: any UK portfolio still reconciling bills against estimated or non-half-hourly data has under a year before that basis disappears. Settlement granularity is also what makes the Australian story above visible in the first place.
Global demand growth is re-accelerating. The IEA now expects world electricity demand to grow 3.6% this year and 3.8% next, up from 3% in 2025, with renewables growing more than 8% in 2026 and reaching 37% of the mix by 2027. So what: rising demand alongside a faster renewables build is exactly the combination that produces flatter averages with sharper edges, which is a good reason not to read one cheap quarter as a durable trend.
Data centres are competing for the solar sponge. NSW data-centre demand averaged 451 MW last quarter, up 35% year on year, and Victoria's nearly doubled from 104 MW to 204 MW (AEMO). AEMO credits growing industrial and data-centre load, alongside home battery charging, with lifting daytime operational demand enough to offset the growth in distributed solar. So what: daytime prices held flat this quarter while every other period fell, so midday load-shifting business cases deserve re-testing rather than rolling forward on last year's assumptions.
From the blog
If you're re-pricing the peak component of a contract, our guide to tariff optimisation covers the mechanics underneath it: how NUOS charges, time-of-use windows and demand tariffs interact, and where the recoverable money usually sits. Read it here.
The Sign-Off
The question I keep coming back to: if batteries have flattened the evening, does peak demand management still earn its place in a commercial energy strategy? My honest read is yes, but for a different reason than it used to have. The energy arbitrage case is thinner than it was. The network demand charge case hasn't moved at all, because that one is set by your own site's peak rather than the market's. Those two arguments have been bundled together for years, and this is the quarter they came apart.
If you've re-run a load profile against the new shape and found something I've missed, just reply. I read every one.
Until next fortnight,
Cohen
Find me on LinkedIn.
Get Behind the Meter in your inbox every fortnight. Subscribe →