At 4:18pm on 28 September 2016, a severe storm system spawned at least two tornadoes across South Australia's mid-north, tearing down 23 transmission towers in minutes. Within seconds, the cascading loss of generation tripped the Heywood interconnector to Victoria, and the entire state — 1.7 million people (approx.), every hospital, traffic light and water pump — went black. Adelaide's hospitals ran on diesel generators through the night; some regional towns stayed dark for days. The Australian Energy Market Operator's (AEMO) final report into the event, released in March 2017, found the storm was the trigger but that protection settings on wind farms turned a survivable fault into a statewide collapse. That night happened under a previous Labor government, Jay Weatherill's, and it is the shadow every South Australian energy minister since has worked in. It is also the reason the state's current push toward a fully renewable grid is not simply a climate story — it is a reliability story, being tested in real time on the same grid that failed once already.
What the government did
Peter Malinauskas's Labor government, elected in March 2022, did not start South Australia's renewables build — it inherited a trajectory running back two decades, through Labor premiers Mike Rann and Weatherill and the Liberal government of Steven Marshall (2018–2022), which legislated a "net 100 per cent renewables by 2030" target. What Malinauskas Labor owns is the acceleration and the plumbing. The government has continued to back Project EnergyConnect, the new high-capacity transmission interconnector to New South Wales being built by transmission operators ElectraNet and TransGrid, as the second lifeline the state needs to trade power with the mainland grid. It committed roughly $593 million (approx., 2022 announcement) to a state-owned hydrogen electrolyser and power station near Whyalla, intended to supply dispatchable, renewables-derived backup when wind and solar fall away. It has continued the rollout of grid-scale batteries and system-strength infrastructure — including ElectraNet's four synchronous condensers, large spinning machines that mimic the physical inertia a coal or gas turbine once provided for free — that AEMO identified as necessary once conventional generators became a minority of what keeps the grid stable. And in 2024, the Department for Energy and Mining's own modelling pointed to South Australia reaching net 100 per cent renewable generation by 2027 (approx.), pulling the inherited 2030 target forward by three years. Once complete, Project EnergyConnect is expected to add up to roughly 800MW of interconnection capacity between South Australia and New South Wales (approx.), close to doubling the state's total capacity to trade power with the rest of the National Electricity Market — a scale of insurance the pre-2016 grid never had. The Whyalla hydrogen plan, unveiled by Premier Malinauskas and then-Energy Minister Tom Koutsantonis in September 2022, envisaged a roughly 250MW hydrogen-fuelled power station (approx.) alongside a large electrolyser, designed to store surplus renewable energy as hydrogen and burn it for power when wind and solar output is low — in effect, a long-duration substitute for the hours and days a battery cannot cover.
The strongest case
South Australia is a genuine world first, and it has done the hard, unglamorous engineering to back it up
No comparably sized grid anywhere has run this much of its electricity from wind and solar this consistently. South Australia's last coal-fired power station, Northern at Port Augusta, closed in May 2016, and the state has not burned coal for power since. AEMO's Quarterly Energy Dynamics reporting has repeatedly shown South Australia's wind and solar meeting more than 70 per cent of local generation over a full quarter (approx., 2023–2024), with periods where renewable output exceeds the entire state's demand and the surplus is exported. This was not achieved by wishing the reliability problem away. After the 2016 blackout, the state government of the day contracted the Hornsdale Power Reserve — the "Tesla big battery," expanded in 2020 to 150MW/193.5MWh — which proved batteries could arrest frequency drops faster than any conventional generator, a result independently verified by AEMO and widely credited with cutting the cost of frequency-control services in the National Electricity Market. The Malinauskas Government has kept building on that foundation rather than retreating from it: more transmission capacity via Project EnergyConnect, more system strength via synchronous condensers, and a genuine attempt at dispatchable, renewables-based backup via the Whyalla hydrogen plan. Grid operators and engineers from Texas to Germany now study South Australia specifically because it is the live experiment the rest of the world has not yet had to run.
The evidence
The headline share of renewable generation is the easiest number to find and the easiest to misread. AEMO's data show the trajectory is real and accelerating — from a negligible share in the mid-2000s to roughly 40 per cent around the time of the 2016 blackout, past 60 per cent by 2020, and above 70 per cent of local generation by 2023–2024 (all approx., AEMO Quarterly Energy Dynamics and annual reporting). But "share of local generation" is not the same as "always self-sufficient." South Australia remains a single small grid at the end of the National Electricity Market, and its ability to run at very high renewable shares depends on being able to both export surplus wind and solar when it has too much, and import power — often coal- or gas-generated in Victoria — when it does not have enough. The Heywood interconnector (upgraded to roughly 650MW capacity, approx.) has carried that load for years; Project EnergyConnect is being built specifically because AEMO's Integrated System Plan judged one interconnector insufficient insurance for a grid this renewables-heavy.
There is a genuine efficiency story inside these numbers too. The Australian Energy Regulator and AEMC have both found that South Australian wholesale electricity prices, and the cost of the frequency-control services that keep the grid's frequency stable, fell substantially after Hornsdale Power Reserve began operating in 2017 (approx.) — a grid-scale battery can respond to a frequency deviation in a fraction of a second, far faster than a gas turbine can start. That is a real, measurable benefit of the renewables-plus-battery model, and it is one reason the rest of the National Electricity Market has followed South Australia's lead on grid-scale batteries rather than the other way around.
The critique
The gap between the renewables share South Australians are told about and the engineering bill required to keep that grid stable is where the government's story gets thinner. Synchronous condensers, batteries, a second interconnector and a hydrogen power station are not free, and none of them existed in the pre-2016 grid. ElectraNet's four synchronous condensers cost roughly $166 million (approx.) to build. Project EnergyConnect's price has grown from an original estimate near $1.5 billion (approx., 2021) to more than $3 billion (approx., 2024 revision), with completion pushed from 2024 toward 2026 (approx.) — a cost that ultimately lands on transmission charges paid by every household bill. These are the visible costs of "firming" a renewables-heavy grid, and South Australian households already carry them: the Australian Energy Market Commission's annual Residential Electricity Price Trends reporting has for years placed South Australia among the highest average retail electricity prices of any NEM state, a ranking that predates and persists through the renewables transition. ESCOSA's own market monitoring has noted that network costs — the poles, wires, synchronous condensers and interconnectors needed to keep a renewables-heavy grid stable — make up a rising share of the typical South Australian power bill even as wholesale energy costs have fallen (approx., ESCOSA Annual Performance Report, 2023–2024).
The dependence on the wider grid cuts both ways, and it is easy to miss inside a headline renewables share. On days of strong wind and sun, South Australia exports its surplus north; on days of weak wind and heavy cloud, it imports power through Heywood from a Victorian grid still substantially fuelled by brown coal, and it still leans on gas-fired capacity at Torrens Island and Pelican Point to cover the gap AEMO's reliability standard requires. None of that shows up in a 72 per cent renewables-share figure, because that figure measures generation, not the imports and gas-fired backup the state draws on to stay reliable through the troughs. A state marketing itself as approaching 100 per cent renewable is, on its lower-wind days, still meaningfully dependent on a fossil-fuelled neighbour and its own gas peaking plants — a nuance the government's messaging rarely volunteers.
South Australia has proven wind and solar can run a grid most of the time. It has not yet proven they can run it cheaply, or without a second grid standing behind them.
The government's own flagship attempt at dispatchable, renewables-derived backup has struggled. The $593 million (approx.) Whyalla hydrogen electrolyser and power station, announced in September 2022 as the answer to "what backs up the battery when the battery runs out," has slipped well behind its original schedule. The wider bet on Whyalla as a green-industry hub was further clouded when the state government placed the associated Whyalla Steelworks, owned by GFG Alliance, into administration in February 2025 (approx.) — a separate crisis, but one that leaves a meaningful share of the government's post-coal energy strategy tied to the fortunes of a single regional project rather than a diversified fleet of dispatchable capacity. Meanwhile, AEMO's Electricity Statement of Opportunities has repeatedly flagged the years before Project EnergyConnect's completion as carrying reliability risk for South Australia, should a summer heatwave coincide with low wind output and an outage on the single existing interconnector or a major gas plant (approx., AEMO ESOO, 2023–2024 cycles). None of this means the renewables share is fake. It means the 72 per cent figure the government cites is only half the sentence — the other half is what it costs, in dollars and in exposure, to keep the lights on the 28 per cent of the time wind and solar are not enough.
The risk to the future
The risk to the future
A grid with almost no room for a second mistake
South Australia is asking a smaller, more isolated grid to do something no comparable grid has sustained for long: run on close to 100 per cent variable renewable generation while staying reliable for 1.8 million people (approx.) and the hospitals, industry and homes that cannot tolerate another 2016. The insurance for that bet — Project EnergyConnect, the synchronous condensers, the batteries, the Whyalla hydrogen plant — is real, but it is also expensive, still incomplete, and partly dependent on a single struggling regional project. If EnergyConnect's delays extend further, if the hydrogen plan does not deliver dispatchable capacity at scale, or if a severe weather event tests the system before the insurance is fully in place, South Australia's genuine global leadership on renewables risks becoming the case study for what happens when the firming does not keep pace with the target — not the case study for a clean transition done right. AEMO's formal reliability standard allows no more than 0.002 per cent of annual demand to go unserved in a year; meeting that standard on a grid this reliant on wind, solar, two interconnectors and one still-unfinished hydrogen plant leaves less margin for error than the coal-and-gas grid it replaced, precisely because there are more moving parts that all have to work at once. The next stress test of this grid will not care which party is in office when it arrives; it will only test whether the wires, the batteries and the backup generation are actually finished.