Around Lake Macquarie, “what happens when Eraring shuts” usually arrives wrapped in the phrase “reliability cliff.” The phrase does not appear in the documents. What the documents contain is more specific, and worth setting out precisely, starting with the housekeeping problem that different AEMO publications are frozen at different moments.
First, the dates, attributed
| Document | States Eraring closes | Vintage |
|---|---|---|
| Origin, extension announcement | April 2029 | 20 Jan 2026; the operative date |
| AEMO, 2026 Integrated System Plan | April 2029, “incorporated into the ISP assessment” (p. 28) | 25 Jun 2026; current |
| AEMO, 2025 Electricity Statement of Opportunities | 19 August 2027, “as advised” by the operator, used throughout its modelling (p. 5) | Aug 2025; finalised on information as at 1 Jul 2025, before the extension |
| Transgrid, Hunter Transmission Project page | “planned 2025 closure” | undated page copy, still live 18 Jul 2026 |
None of this is scandal; documents freeze when they are published. But it means any claim that starts “AEMO says...” about Eraring needs a vintage attached, and it means the 2025 outlook’s Eraring-specific modelling ran on a date that is now 20 months early. Keep that in mind for everything below; it mostly makes the findings stronger, not weaker.
What the reliability outlook actually found
The Electricity Statement of Opportunities is AEMO’s annual ten-year reliability forecast. It measures reliability against a standard: expected unserved energy, the fraction of demand that might go unmet, must stay below 0.002 per cent per region per year, with an even stricter interim measure applying until mid-2028 (pp. 5, 9). The 2025 edition’s sentence that matters most to this region is on page 4:
“Reliability gaps previously identified following the advised retirement of Eraring Power Station in 2027-28 in New South Wales are no longer forecast. However, ... system security and operability challenges remain.”
Unpack that. Modelling a 2,880 MW station leaving in August 2027, the earlier date, AEMO found NSW no longer faces a forecast breach of the reliability standard at the exit. Under its stricter “Committed and Anticipated Developments” scenario, which only counts projects meeting formal commitment criteria, the outlook’s forecast gaps sit elsewhere: an 80 MW gap in Queensland in 2025-26, a 390 MW gap in South Australia in 2026-27, then nothing in NSW until 2033-34 (pp. 9-10, Figure 3). And the NSW gap, when it comes, has a named trigger, one with a Hunter address: the outlook attributes it to the advised 2033 retirements of Bayswater (2,173 MW) and Vales Point (1,320 MW), compounded by wider fleet exits and demand growth (p. 61), not to the specific effect of Eraring leaving. And since the station will now run 20 months longer than that modelling assumed, the near-term picture can only have more supply in place at the exit than the model saw, though we note that inference is ours; the re-run on the new date belongs to the 2026 outlook, due around August.
The same document’s caveat deserves equal weight, because it is the half the “cliff” shorthand gets at. AEMO distinguishes capacity reliability, having enough megawatts across a year, from system security: keeping voltage, frequency and system strength stable minute to minute, services large spinning generators like Eraring currently provide. On security and operability the outlook says challenges remain, and points to further system-security analysis. That is the precise, checkable worry, and it is a different claim from “the lights go out in 2029.”
What the current plan adds
The 2026 Integrated System Plan, the up-to-date document, incorporates the April 2029 date and sets the wider context: of the NEM’s 26 major coal stations in 2012, ten have already closed; all but one of the remainder have announced closure dates, more than half by 2035; and AEMO’s modelled least-cost path retires nearly two-thirds of the remaining fleet by 2035, with only Queensland coal operating beyond 2038 (p. 76). On the supply side it counts the batteries, zone upgrades and transmission projects this masthead tracks individually, plus the Hunter Power Station gas peaker at Kurri Kurri, whose second unit the market operator’s quarterly data records as having begun commissioning in December 2025.
So the state of the paperwork, dated and attributed, is this: the operative closure date is Origin’s April 2029; the current system plan already assumes it; the most recent reliability modelling, run on an earlier date, found no NSW capacity gap at the exit and a late-decade gap driven by the fleet, not the station; and the open technical question AEMO itself flags is system security, not megawatts. The 2026 Statement of Opportunities will be the first reliability run on the real date. When it publishes, we will read it and report what changed.