Scenario architecture: four drivers decide whether 2030 lands nearer 5 GW or 6.2 GW, and the pipeline sets the upper end
Four drivers decide whether 2030 lands nearer 5 GW or 6.2 GW. The bear, base and bull cases, what each assumes, and why even the bear case is a roughly 3.7x expansion on end-2025 operating capacity.
Operating capacity, GW: published figures above the line, the inashanu 2030 scenario corridor below
Four drivers decide where in the corridor 2030 lands
Power access & policy
AI demand conversion
Cost of capital
Delivery capacity
Anchors (all published)
| Anchor | 2030-ish figure | What it counts |
|---|---|---|
| AEMO Step Change AEMO 2025 IASR | ~6 GW required | Operating IT load |
| Mordor capacity forecast Mordor (Jan 2026) | 7.18 GW (15.25% CAGR) | IT load capacity |
The base case is the AEMO requirement met. Mordor is carried on its own basis. M3 (~9.5 GW, early 2030s) and CEFC/Baringa (4.7-7.4 GW, 2035) follow overleaf.
All three scenarios accept the demand evidence (Sections 04, 08): they differ on deliverability and financing, not on whether customers want Australian megawatts
The four drivers: bear and bull conditions
| Driver | Bear condition | Bull condition |
|---|---|---|
| Power access & policy | Net-generator costs + NSW prohibitions stall connections; queue stays 8+ yrs | REZ-style access schemes accelerate; firming obligations met with BESS at scale |
| AI demand conversion | OpenAI-scale MoUs lapse; inference shifts to edge; US-style construction rollover arrives | MoUs convert to leases; sovereign-AI exemption pulls regional GPU demand to AU |
| Cost of capital | Rates higher-for-longer; NXT-style de-rating spreads; private debt tightens | REIT IPO window opens (AirTrunk); superannuation allocations deepen |
| Delivery capacity | Trades competition with housing worsens; 22% senior engineer vacancy persists | IDA fast-tracking holds; modular/prefab lifts throughput |
How to use these scenarios
They are planning constructs, not forecasts: each pairs a capacity outcome with the driver conditions that produce it, so a committee can map its own judgement to a number. The probability weights on the watchboard page make our judgement explicit; the driver table above shows what evidence would change it.
Anchors (all published), continued
| Anchor | 2030-ish figure |
|---|---|
| M3 built-out trajectory | ~9.5 GW by early 2030s |
| CEFC/Baringa | 4.7-7.4 GW by 2035 |
| M3 supply gap warning | up to 1.7 GW gap even after doubling |
Anchors (all published), continued. Source: AEMO 2025 IASR; Mordor (Jan 2026); M3 (Nov 2025); ITK/CEFC-Baringa (May 2026). As at 21 Sep 2026.
- Bear 5.0 GW
- Low end of the 5-8 GW planning band (dossier synthesis, 17 Sep 2026)
- Base 6.0 GW
- The AEMO Step Change requirement met in full (2025 IASR)
- Bull 6.22 GW
- Every under-construction megawatt commissions, plus 35% of planned and deferred capacity (project table, Section 05)
Base case path: ~6.0 GW operating by 2030, a 4.4x expansion in five years, with the bull only 220 MW above it
Three capacity paths from a common 1.35 GW base, GW
Bear 5.0 GW
Base 6.0 GW
Bull 6.22 GW
Base-case delivery path
Committed UC delivers (CDC EC5/6 +108 MW, Laverton first campus 150 MW, Stack MEL01/02), the low-risk tranche; AirTrunk MEL2 targets opening in 2H 2027.
CDC's 555 MW contract begins delivering (FY28-29); Marsden Park phases; S7 Phase One (H2 2027 target) if the MoU firms.
Second-wave campuses (SYD4, ISPT, Atlas, Horsley Park) deliver into net-generator-compliant power.
What moves the path off base
What moves the path off base. Source: inashanu estimates. As at 21 Sep 2026.
- Bear trigger
- First ministerial network-access prohibition, or OpenAI MoU lapsing, takes the path toward 5 GW.
- Bull trigger
- The whole under-construction pool commissioning on schedule, plus a second 500 MW-class contract, takes it to 6.22 GW and no further on announced projects.
Why not simply Mordor's 7.18 GW?
Mordor counts IT load capacity: 3,530 MW in 2025 (Jan 2026), against 1,350 MW of operating IT load here (ITK, 11 May 2026). A rate between those two ends measures the distance between two definitions as well as five years of building. At that ratio, 38.2%, its 2030 figure restates to about 2,746 MW (Section 02).
Why the bull sits so close to the base
The bull is built from the projects that exist: 2,301 MW under construction in full, plus 35% of the 5,892 MW planned and of the 1,450 MW Firmus Southgate balance (Section 05, to 16 Sep 2026). That reaches 6,220 MW, 220 MW above base. The deliverable pipeline carries no dramatic upside; more needs unannounced projects on unallocated power, and neither is an 18-month matter.
Even the bear case is a growth story: 5.0 GW by 2030 is ~3.7x end-2025 operating capacity at ~30% p.a.
First authorisation/prohibition
The single most powerful siting signal in the market
inashanu probability weights (judgemental)
Bear ~25%
Base ~55%
Bull ~20%
inashanu probability weights (judgemental). Source: inashanu synthesis. As at 21 Sep 2026.
- Raises floor
- Demand evidence is strong
- Caps ceiling
- Deliverability screens remove most of the announced pipeline
- Widens tails
- Policy is mid-rewrite
Weights are committee-discussion inputs; no model produces them.
What the scenario debate is really about
The scenario debate is about the slope of extraordinary growth, and about who owns the deliverable megawatts when the music slows.
The scenario set in one panel (inashanu estimates)
Sources
- AEMO 2025 IASR
- Mordor, as at Jan 2026
- M3, as at Nov 2025
- ITK/CEFC-Baringa, as at May 2026
- inashanu estimates
- anchors: ITK (May 2026), AEMO 2025 IASR, Mordor (Jan 2026), M3, as at Nov 2025
- inashanu synthesis
Related
- What will the Australian data centre market be worth in 2030?
- How much of Australia's announced data centre pipeline will actually be built?
- How much electricity do Australian data centres use?
- What is in Australia's data centre development pipeline?
The full primer is published as an 85-page PDF (download) with the underlying market model as a workbook (download).
This page is one of 30 in Australian data centres: the research, indexed by the question each one answers, with the four pages where the published sources disagree set first.
The research is published by inashanu, a studio that produces fixed-scope market research for Australian private equity, priced and dated before work starts. The sample engagement is this primer in full.
This standard, on your question. Six fixed-price deliverables, A$14,000 to A$45,000 excluding GST, opening at the A$14,000 market map in 10 business days. State the decision and the date it has to be made; the written assessment of whether the evidence base will carry it follows within two business days and carries no charge. Request a brief assessment · the catalogue and its prices.