Just after 12:30pm on a clear Saturday in late October, operational demand on the South Australian grid dipped below zero. Not low – below zero. AEMO’s own data on minimum demand events shows this isn’t a one-off curiosity anymore; it’s a recurring feature of a mild, sunny weekend with everyone at the beach or the footy and panels doing exactly what they were installed to do. I was out at a client’s place in the Adelaide Hills that same week, checking a system she’d had in eighteen months, and her inverter was happily throttling itself back mid-afternoon with nobody home to see it.
That throttling is the whole story in miniature. Rooftop solar’s success has created a problem that none of the original installers, policymakers or even AEMO fully war-gamed a decade ago: too much of exactly the right kind of generation at exactly the time nobody needs it.
What minimum demand actually means #
Minimum demand isn’t about how much power the grid is using. It’s about how much power the grid needs to draw from the transmission system, once you net off everything households and businesses are already generating themselves behind the meter. When rooftop solar output is high and underlying consumption is low – think a mild autumn Sunday, washing machines off, air conditioners off, everyone out – the grid-supplied portion of demand can fall to levels that used to be unthinkable even at 3am.
Think of it like a river with lots of small creeks feeding in upstream. If every creek is suddenly running at flood level at the same time the main channel barely needs any water, the river authority has to find somewhere for that extra flow to go, or slow the creeks down. Rooftop solar is the creeks. The NEM operator is the river authority, and for a long time it didn’t have many levers to slow the creeks down gracefully.
Why midday is the flashpoint, not peak demand #
For most of the grid’s history, the hard planning problem was the opposite one: the hot January afternoon when every air conditioner in Sydney switches on at once and the system strains to keep up. That’s still real, and still drives a lot of network investment. But the midday problem is a newer, quieter kind of stress. It’s not about running out of supply. It’s about having too much of one type of supply concentrated in a narrow band of hours, with very little flexible load or storage sitting between the panel and the transmission network to soak it up.
South Australia has been the testing ground, as it tends to be – I’ve written before about how South Australia keeps teaching the NEM lessons it didn’t ask for. Queensland and NSW are catching up fast on rooftop capacity, and both AEMO’s Integrated System Plan and the state-level data point to the same pattern spreading east as panel numbers climb past four million installations nationally, per the Clean Energy Regulator’s small-scale technology data.
The voltage problem hiding inside the demand problem #
Here’s the bit that doesn’t get explained well in the glossy brochures: minimum demand isn’t just a supply-and-demand balancing act, it’s a voltage problem on your own street. Distribution networks were built on the assumption that power flows one way: down from the substation, out along the feeder, into your switchboard. Push enough solar back up that same feeder at midday and voltage starts climbing at the far end of the line, often right where a cul-de-sac of north-facing roofs has every house exporting at once.
That’s why your installer size-checks string lengths and cable runs so carefully, and why export limits exist at all. A 5kW system with a 1.5kW export cap isn’t being stingy for no reason: the distributor is managing a voltage envelope under AS/NZS 4777, the standard that governs how inverters connect to the grid, and your particular feeder might simply not have headroom for more. I’ve had customers ring up convinced their installer short-changed them on export limits, and sometimes that’s a fair complaint. More often it’s a network constraint that has nothing to do with the install and everything to do with how many other rooftops are already feeding that same transformer.
What AEMO and the networks are actually doing about it #
The practical response has been a mix of market mechanisms and blunt engineering. AEMO can issue minimum system load instructions and work with distributors on ways to manage the risk directly, including the dynamic export limits now rolling out in parts of the network: inverters that can be told, in near real time, to turn exports up or down depending on local conditions rather than being locked to a single static cap forever. On the generation side, utility-scale batteries are increasingly being built with an eye to absorbing exactly this midday glut; I’d point to the way projects like the Waratah Super Battery is earning its keep as a shock absorber for grid stress, even though its primary job description is frequency and system strength rather than daytime solar soak.
Grid-forming inverter technology matters here too, and it’s worth understanding why. I went through the mechanics in a separate piece on why the grid suddenly needs grid-forming inverters. The short version for this topic: as big, spinning coal and gas units retire, the grid loses some of its natural inertia and voltage-steadying behaviour, right at the same time rooftop solar is pushing more power the wrong way down the wires. Both problems land on the same stretch of the day.
What this actually means for a household weighing up solar or a battery #
On the numbers, here’s what a real household in a solar-saturated postcode sees today. Your export limit might be lower than your neighbour’s who installed three years earlier on a less congested feeder – pure bad luck of timing and location, not a reflection of your system’s quality. Feed-in tariffs for midday exports have been sliding for years and several retailers now run time-varying tariffs that pay close to nothing, or occasionally charge you, for power pushed into the grid between 10am and 3pm, while paying considerably more for exports in the early evening ramp. The AER’s default market offer determination each year increasingly reflects this shift in the underlying wholesale shape, something I touched on when writing about the Default Market Offer and why it keeps rising.
That’s the case for sizing a system to your own consumption rather than to the biggest roof you can fit, and increasingly the case for pairing solar with a battery that shifts your own midday glut into the evening, rather than selling it cheap and buying it back dear six hours later. I ran the actual payback maths on that trade-off properly in an earlier piece – home battery payback: does it stack up yet, and the answer isn’t a flat yes or no, it depends heavily on your export limit and your evening usage pattern.
A system that was built for scarcity, meeting an age of abundance #
My mildly unpopular view on this, and I’ll own it: a decent chunk of the industry spent the 2010s selling rooftop solar as a straightforward bill-reduction product, when really it was always going to become a grid-participation product with constraints attached. Nobody wanted to say that to a customer standing in their driveway holding a brochure with a 6-year payback headline on it. We’re now retrofitting the honesty into the conversation a decade late, via export limits and flat feed-in tariffs that feel, to a lot of homeowners, like the rules changed after they bought the ticket.
They didn’t really change the rules so much as catch up to a reality nobody modelled properly the first time around: a bit like how I kept killing the same basil plant on the north windowsill for two summers before realising the problem wasn’t the plant, it was that I’d put six of them in a pot built for two. Four million-plus rooftop systems feeding into a grid built for one-way power flow was always going to hit a ceiling somewhere. The question now isn’t whether that ceiling is real, AEMO’s own minimum demand figures settle that, it’s how fast batteries, dynamic export limits and genuine demand flexibility arrive to lift it before rooftop solar’s growth curve starts bending under its own success. For a deeper look at the broader tension, my colleague’s piece asking whether rooftop solar is quietly breaking the grid is worth sitting with.
Whether your own system hits that ceiling sooner or later mostly comes down to which feeder you’re sitting on and how many of your neighbours got in before you. Worth checking with your distributor before you sign anything, not after.
– Priya Nair, Solar & Distributed Energy Correspondent
Photo by Nuno Marques on Unsplash