The Fragile Promise of a Net-Zero Grid

Spain’s 2025 blackout exposed the fragility of inverter-dominated grids. Australia’s net-zero plan must master four physical pillars – energy, capacity, inertia, absorptivity – or risk abundance becoming the next blackout.

Spain’s Wake-Up Call

At 12:33 p.m. on 28 April 2025, the Iberian Peninsula went dark. A cascading voltage surge swept through the Spanish and Portuguese grids, tripping inverters and transmission lines in rapid succession. Within seconds, generation collapsed, cutting power to tens of millions of people.

There was no cyberattack. No storm. Just sunshine – and too much of it. Mid-day solar output was enormous, conventional units were idling, and the system had too little grid inertia to absorb the surplus. Voltage protection relays triggered, inverter logic misbehaved, and the grid collapsed. Power was restored within hours, but the technical investigation revealed systemic vulnerabilities that extend far beyond the Iberian Peninsula – weaknesses common to every grid now dominated by inverters rather than turbines.

The episode was the largest power outage in modern European history and a warning to every renewable-powered grid: stability can fail from abundance just as easily as from shortage. The detailed technical sequence remains under review, but the broad lesson is already clear – a grid rich in renewables must preserve the physical properties that once came free with conventional generation.

Australia is watching closely, because its own plan heads in the same direction. By the 2040s, almost all coal-fired synchronous generation in the National Electricity Market (NEM) will be retired. Solar and wind will supply nearly all energy; batteries, pumped hydro, and interconnectors will balance supply; gas will recede to emergency use. 

Australia's is a bold plan – but also a fragile one. Coal plants don’t just generate electricity; they provide capacity (instant power), energy depth (multi-day endurance), and inertia (frequency stability). They can also be ramped up and down to match demand – a controllability that renewable generation lacks without large-scale storage or flexible load. Remove them, and those functions must be replicated by new technologies – or reliability will fray.

The Physics of Reliability: the four foundations

Every renewable-powered grid must satisfy four physical conditions simultaneously: energy, capacity, inertia, and absorptivity. These are not optional features of a clean system; they are the foundations that keep it stable, controllable, and secure. Australia’s net-zero pathway must clear all four hurdles at once.

Energy – Multi-day storage (GWh)

The first limit is energy depth – the stored power needed for multi-day lulls in wind and solar. Europe calls this Dunkelflaute – “dark lull.” Australia’s own climate brings calm, clouded weather systems that can linger for days. Average daily NEM demand is about 525 GWh; AEMO’s 2024 Integrated System Plan calls for ≈ 650 GWh of dispatchable storage by 2050 – roughly one day’s national consumption. Today, the grid has about 5 GWh of large-scale, grid-connected batteries, most designed for one- to four-hour discharge. Projects under construction – such as Waratah Super Battery, Torrens Island 2, and Hazelwood – will double that within two years, but it remains small beside total demand. 

Behind the meter, the private and household sector now contributes an estimated 15–20 GWh of installed storage – roughly 10–12 GWh in household battery systems and 5–8 GWh in electric vehicles. Yet almost none of this capacity is dispatchable by the market operator: household systems are optimised for self-consumption, and EV batteries remain largely untapped. Until these assets can be aggregated or dispatched through virtual power plants or vehicle-to-grid networks, most of Australia’s storage will remain outside AEMO’s control.

Snowy 2.0 is the cornerstone: 350 GWh of storage energy, but only 2.2 GW can flow at once – about 6 per cent of NEM peak demand. It would take a week to empty at full load. Snowy therefore contributes depth but not breadth: valuable endurance, limited reach. A three-day calm could leave a 1.5 TWh deficit – four Snowys’ worth of stored energy.

Capacity – Peak power delivery (GW)

The second limit is peak delivery capacity – the ability to meet peaks after sunset. NEM peaks around 33–35 GW. Coal still supplies about half. Current firming assets – hydro, Snowy 2.0, and batteries – provide only about 10 GW combined. To replace coal, Australia will need another 25–30 GW of firm power: batteries, hydro, gas, or new technologies. A system can hold energy yet still fail if it cannot release enough power at once – the distinction between stored energy and deliverable capacity lies at the heart of grid reliability.

Inertia – Frequency stability (cycles per second)

A third limit is inertia, the hidden foundation of grid stability. Frequency is the primary expression of the balance between generation and demand, measured in cycles per second (hertz). Power in alternating current has three interlocking components: real power (P), which does useful work; reactive power (Q), which sustains the oscillating electric and magnetic fields and makes that useful work possible; and apparent power (S), the total flow that combines them. When that triangle slips out of balance, voltage wobbles and frequency soon follows.

The old turbines held that balance effortlessly. Their great steel rotors carried the weight of the system in their momentum. Each generator was both muscle and metronome, its magnetic field binding distant parts of the grid into a single pulse. Inverter-based generation has no such mass; its control is digital, not mechanical, and the coordination between power, phase, and frequency – what operators call the dance of volts, vars, and hertz – can be fragile. When inertia is thin, frequency can swing several hertz per second, too fast for protection or human control to catch.

Spain’s 2025 blackout showed how low inertia and weak voltage support can turn a local disturbance into a continental collapse. South Australia already runs on roughly one-tenth of its historical inertia; synchronous condensers and grid-forming inverters help, but the margin is slender. Inertia is measured not in megawatts but in seconds – the breathing space between a disturbance and a blackout.

Absorptivity – Surplus management (flexible GW)

Finally comes absorptivity – the ability to curtail, soak up or shed surplus power before it destabilises the system. On bright, windy days, Australia already produces more electricity than demand for hours at a time. In conventional systems, heavy generators absorb the excess; in inverter-dominated ones, protective logic can trigger mass trips. Spain’s blackout began from abundance, not shortage. 

By 2030, rooftop exports will frequently exceed daytime demand. Without flexible loads, the system will oscillate between glut and scarcity.

Managing this surplus requires flexibility – the ability to use, redirect, or curtail power safely: dynamic tariffs to shift demand; flexible industrial users such as hydrogen or desalination; large-scale storage; and, crucially, markets that pay for curtailment headroom as a stability service. 

Absorptivity is the demand-side twin of storage. Storage shifts energy through time; absorptivity removes excess in the moment. Without it, the grid becomes a seesaw of feast and famine.

Bridging these gaps will take investment on the scale of the original Snowy Mountains Scheme – perhaps several of them – plus a digital revolution in demand-side control.

Progress and Pressure

Not everything is bleak. Battery deployment is accelerating sharply. Projects such as Torrens Island, Waratah Super Battery, and Hazelwood are adding multiple GWh each year. Grid-forming inverters are maturing; South Australia can now operate briefly in 100 per cent renewable mode. AEMO’s forecasting accuracy has improved; short-term solar and wind predictions are now far more reliable. 

Virtual-power-plant trials show real promise. South Australia’s VPP program now aggregates more than 4,000 home batteries with a combined capacity of around 40 MWh, providing frequency control and contingency services that once required dedicated generators. Similar pilots in the ACT and Victoria are expanding rapidly, proving that distributed storage can operate as genuine grid infrastructure when coordinated effectively.

Planning has also improved since the 2016 South Australian blackout. System-restart procedures, protected-operation modes, and inertia-monitoring tools now exist. But the system still assumes storage and diversity will always suffice. Spain’s failure shows that rare alignments of stress – high renewables, low inertia, weak voltage support – can trigger cascades faster than models predict. Australia, isolated from any larger grid, must plan for self-restoration in the dark.

Yet both demand-side reform and climate adaptation lag behind the pace of generation investment. A 10 per cent fall in total demand – through efficiency, insulation, or smarter timing – would ease every one of the four constraints. Yet efficiency rarely attracts the political glamour of new generation. 

Meanwhile, hotter temperatures reduce transmission capacity and thermal generation efficiency while raising cooling loads. The NEM’s physical headroom literally shrinks on the days it is most needed.

Markets, Politics, and the Path Forward

Building 650 GWh of storage and 30 GW of firming will cost tens of billions over the next two decades - roughly comparable to building several new Snowy schemes, the original having cost approximately $8 billion in today's dollars.

Yet an energy-only market under-rewards assets that rarely run. Spain made the same error: too few voltage-supporting units online because the price signal was weak. Australia’s new capacity and inertia markets are a start, but incentives remain patchy. Without clear value for flexibility and stability, investment will lag – and governments will be forced to intervene, often by keeping ageing coal units online longer than planned.

Interconnection helps but cannot save. EnergyConnect and Marinus Link will improve diversity, but they cannot conjure wind or sun. When continental weather systems calm, every state slumps together. Interconnectors move power; they do not create it. The NEM must be self-reliant.

Some argue for brute-force over-building: install more solar and wind than needed, accept heavy curtailment, and channel the surplus into hydrogen or industry. That can reduce the need for long-duration storage, but it raises cost, land-use, and operational complexity. Industrial off-takers cannot always respond fast enough to stabilise frequency or voltage, and hydrogen demand itself depends on market economics. Spain’s lesson still applies: over-abundance without controllable absorptivity remains a source of instability.

AEMO still assumes a small gas reserve as strategic insurance. Run rarely, these turbines emit little but buy priceless resilience. Until long-duration storage or hydrogen turbines are proven, maintaining a modest dispatchable fleet is pragmatic, not regressive. The point is not nostalgia for coal or gas – it is respect for arithmetic.

But the barriers are not only technical. Investors seek certainty, incumbents protect sunk assets, and governments fear consumer backlash from rising bills. Each delay in transmission approval or pricing reform benefits the status quo and pushes costs onto the future system. Political economy, not engineering, may prove the hardest constraint of all.

A Fair but Concerned View

Australia’s net-zero grid plan is visionary but tightly coupled. It assumes that hundreds of GWh of storage, tens of GW of firm capacity, widespread synthetic inertia, and flexible demand will all arrive on schedule, at falling cost, and in perfect coordination. 

That is possible – but not inevitable. To make it real, policy must: 

1. Accelerate large-scale storage and firming investment through clearer market value or public co-funding. 
2. Reform pricing so flexible demand is rewarded as reliability infrastructure. 
3. Sequence coal retirements only when equivalent inertia and dispatchable replacements are physically ready. 
4. Maintain modest firm backup – gas, hydro, or potentially nuclear, despite its political and regulatory challenges in Australia – until multi-day storage is proven. 

Each of these measures demands coordination, accountability, and a willingness to confront physical limits with political honesty. These are not calls to delay decarbonisation, but to deliver it securely. 

The Spanish blackout was a warning shot – not against renewables, but against complacency. It showed how quickly a modern, well-managed grid can unravel when abundance meets fragility. Australia’s challenge is to learn that lesson early. 

A renewable-powered grid can be green and reliable – but only if it holds its four foundations in balance: energy, capacity, inertia, and absorptivity. Ignore any one of them, and the system tilts toward instability. 

Get them right, and Australia’s net-zero promise becomes not just ambitious, but achievable – a model of how to keep the lights on in a decarbonised world.