In 2025-26, 9.1 GW of new generating and storage capacity was connected to the NEM, marking a recent record. The country is closing 15 GW of coal and gas without having adopted a phase-out law, while the NEM has a pipeline of 40 GW of generating and storage projects in development or anticipated. AEMO identifies a least-cost trajectory that integrates public policy, targets and required investments.
The essentials
- AEMO identifies a pipeline of 40 GW of generating and storage projects in development or anticipated, while 15 GW of fossil fuel capacity is scheduled for closure. Investments and retirements are driven by market forces, public programs and coordinated delivery of networks, storage and new capacity.
- The 9.1 GW connected in 2025-26 represents a recent record and reflects new execution capacity across the value chain.
- Electricity demand is expected to grow by 40 percent due to electrification of households, businesses and industry, as well as data center growth: fossil fuels are losing both in production cost and in ability to capture growth.
- The challenge is no longer the start of the transition but its reliability when replacing fossil baseload with variable assets.
- The Australian case tests a thesis: when economic, technological and political forces converge, the shift accelerates and can become irreversible.
Australian fossil fuels lose money before being banned
Australian coal is not closing because a law orders it. Economic pressures are a significant factor in coal retirements, alongside other operational and energy policy factors. A coal-fired plant builds its profitability on hours of production sold at high prices. When solar and wind flood the network at certain times and compress wholesale prices, the hours of high-priced production shrink. Revenue collapses before output does.
This is the mechanism documented by AEMO in its 2026 report: the 15 GW of coal and gas scheduled for closure do not exit at the regulator’s demand. The notified dates rest on information from operators, in an environment where reliability requirements and public interventions can also influence decisions. The state did not have to decide.
This mechanism is not unique to Australia. IRENA has documented for several years the compression of levelized costs for solar and onshore wind below the marginal costs of aging thermal plants in markets exposed to competition. Australia makes it visible because its electricity market is open, its solar and wind resources among the best in the world, and its policy has not imposed a phase-out timeline that could have served as a pretext to delay decisions.
Economic constraint can precede and replace regulatory constraint, provided that the price signal is real and that renewables are competitive enough to transmit it.
40 GW committed, 9.1 GW connected: the transition moves into industrial scale
The doubling of connection rate in one year reflects new execution capacity across the entire value chain: financing, permitting, construction and connection.
The 40 GW represents a pipeline of projects in development and anticipated, not exclusively projects that are financed, contracted or under construction. For a country whose electricity consumption stands at around 200 TWh per year, the transformation of the energy mix depends on the actual delivery of generating, storage, network and consumer resource projects. If demand grows by 40 percent as AEMO projects due to electrification of households, businesses and industry, as well as data center growth, the expected new capacity must contribute to replacing retired capacity and meet demand growth.
Tim Lenton’s thesis on positive tipping points applies here. In his work on socio-technical systems, Lenton identifies moments when the alignment of economic, technological and political forces makes the transition self-sustaining. The system does not reverse because capital is already reallocated, skills already developed and supply chains already oriented toward the winning technology. In Australia, the pipeline reflects generating and storage projects in development or anticipated, whose composition and realization remain to be confirmed.
Minister Chris Bowen called these figures a “record” when announcing them in September 2026. The formula is accurate but incomplete: this record is above all a market signal, not a political achievement.
Reliability, the only variable that can slow the shift
The coordination problem between closures and deployment of flexibility is structurally asymmetrical. A fossil power plant closes in a single administrative decision, often quick once margins turn negative. Storage and transmission lines, meanwhile, follow planning, authorization and construction cycles measured in years. This temporal asymmetry creates a window of vulnerability during which the network loses guaranteed power before acquiring the tools to compensate for renewable variability. The risk is not that the transition fails, but that it produces supply tensions visible enough to fuel political resistance to continuing closures.
Reliability then becomes a political variable as much as a technical one: a major incident, even temporary, can crystallize opposition that would find no other credible argument.
Optimism about commitment figures must not avoid the real problem. Closing 15 GW of fossil baseload—plants that produce on demand, day and night, regardless of weather—and replacing them with solar and wind requires a flexibility infrastructure that Australia is still building.
The Australian network remains fragmented. The network requires some targeted reinforcements to effectively integrate renewable and storage resources. Battery storage projects and new transmission lines are underway, but their deployment pace directly conditions the speed at which fossil closures can be absorbed without risk of failure.
AEMO makes this clear in its 2026 report: system reliability depends on coordination between renewable deployment, storage and network reinforcement. If this coordination fails, in case of reliability risk, the mechanism described by AEMO rests on contractual obligations for retailers, not regulatory postponement of closures. This would be the only circumstance where market logic would meet institutional constraint.
The tension is real. Acemoglu and Johnson, in their work on technology and economic power, remind us that gains from a technological transition are not distributed automatically and that institutions play a decisive role in how risks are borne. In Australia, the situation is concrete: regions hosting coal plants—Hunter Valley in New South Wales and Latrobe Valley in Victoria—depend on these jobs. Market speed can exceed the capacity of just transition policies to manage reconversions. This mismatch does not undermine the direction of the transition, but it conditions its political legitimacy.
Market where law has not acted
Australia has no coal phase-out law. The Albanese government has set a target of 82 percent renewable electricity by 2030 and actively supports projects through the Australian Renewable Energy Agency, but no law imposes closure dates on fossil operators. This legislative gap has become, in practice, an argument in favor of the tipping point thesis.
A phase-out law sets dates that can be poorly calibrated if costs fall faster than expected or if substitute infrastructure is not ready. The market, meanwhile, adjusts continuously. Notified closures rest on information provided by operators, in an environment where reliability requirements, public interventions and the regulatory framework influence decisions.
That said, the market does not solve everything. It does not spontaneously produce the transmission lines the network needs—collective goods that only the state can finance and plan. It does not guarantee just transition for coal basin workers. And it can produce disorderly closures if price signals accelerate faster than replacement infrastructure. Technologies for coal phase-out exist but their large-scale deployment remains a systemic engineering challenge that price signal alone does not solve.
The interest of the Australian case lies precisely here: it shows the conditions under which the market can do much, and delimits what it cannot do alone.
The irreversibility threshold of the shift
The economic irreversibility of a transition rests on capital reallocation but also on skills reallocation. When engineers, contractors and maintenance teams migrate to renewable sectors, fossil plants progressively lose access to the human resources necessary for their safe operation. This phenomenon, distinct from the price signal, reinforces exit dynamics independently of regulatory decisions and tends to accelerate as the share of clean assets in the mix grows. The labor market thus becomes an additional vector of irreversibility that neither fossil operators nor regulators directly control.
Coal exit in Australia is underway. The forward-looking question concerns crossing the point of economic no-return and the lessons other economies can draw from the Australian case.
Lenton and colleagues’ Global Tipping Points Report 2024-2026 identifies several markers of irreversibility in energy transitions: the share of clean energy in new investment, the relative cost of competing technologies, and the dynamics of skills and supply chains. In Australia, these elements point toward growing system reorientation. New fossil capacity is heavily constrained, but AEMO still foresees a back-up role for gas in the NEM trajectory. Engineers and construction firms are turning to renewables. Supply chains are reorienting.
For this shift to reverse would require major exogenous shocks: sustained rare earth price increases, a permitting governance crisis, or political change that would dismantle support mechanisms. None of these scenarios is impossible. But each becomes more costly as investment already made in renewables creates its own economic irreversibilities.
The anticipated rise of over 40 percent in demand increases the need for capacity, network and flexibility; it makes the shift robust only if these investments are delivered on time. Data centers settling in Australia, industrial processes electrifying, vehicles charging: all create growing electricity demand. A portfolio dominated by renewables, connected to the network, reinforced by storage and supported by flexible gas constitutes the least-cost path compared to fossil alternatives. Coal exit results from multiple economic and technical factors, including renewable competition, costs and operator decisions. Demand growth increases the need for new renewable capacity, whose profitability and realization depend on prices, contracts, costs, network access and financing.
This mechanism could interest economies still seeking the right political lever. Gulf oil companies are decarbonizing their production by adjusting extraction costs, but their logic remains fossil. Australia shows another trajectory: let prices do the selecting and build the infrastructure that makes this selection reliable.
The challenge through 2030-2035 is whether flexibility, storage, network and demand management infrastructure will deploy fast enough to absorb closures without supply tension episodes. If yes, the Australian model offers a powerful empirical argument for countries still hesitant between authoritarian regulation and price signal. If no, it will also provide an argument on the limits of market alone.
Both readings would be useful. The transition will not wait for us to choose.
Sources
- AEMO 2026 Electricity Supply and Demand Outlook (ESOO), Australian Energy Market Operator, 2026
- Minister Chris Bowen, “Record clean energy build strengthens reliability outlook”, September 2026
- Tim Lenton et al., Global Tipping Points Report 2024-2026, https://www.globalclimate.org/global-tipping-points-report
- IRENA, Global Renewable Cost Analysis 2026, International Renewable Energy Agency
- Australian Renewable Energy Agency (ARENA), installed capacity data 2025-2026
