5.4 GW of data centres in the connection pipeline
Eleven transmission-connected data-centre projects exceeding 5 MW represented 5.4 GW of maximum demand, concentrated in NSW and Victoria and mostly at an early development stage.
Permanent case linkYour task
Reconcile staged demand against connection limits, compare feasible next steps and identify the evidence needed for an engineering and procurement review.
Your role
Analyst supporting a staged campus connection review
Practice data & worked solution
A staged campus connection
Original synthetic teaching inputs, not an actual developer's plans or AEMO forecast. Each phase is a separate constant-load, full-year scenario of 8,760 hours. The PUE values are power-equivalent approximations for this exercise; real peak demand requires coincident profiles. Energy-only price is A$80/MWh.
| Phase | IT load MW | Assumed PUE | Connection limit MW |
|---|---|---|---|
| Initial | 40 | 1.3 | 60 |
| Expansion | 120 | 1.25 | 150 |
| Full campus | 240 | 1.2 | 250 |
Your analysis
- Calculate facility MW, annual MWh and energy-only annual cost for each phase. State the unit conversions.
- Find which phase exceeds its connection limit. Quantify the shortfall and the largest IT load consistent with that limit under the same PUE approximation.
- Compare a slower IT ramp, a larger connection and a verified flexibility arrangement. Explain why UPS nameplate capacity alone does not solve a persistent connection shortfall.
Worked solution
Load and cost reconciliation
Initial: 40 × 1.3 = 52 MW; 455,520 MWh/year; A$36,441,600/year. Expansion: 150 MW; 1,314,000 MWh/year; A$105,120,000/year. Full campus: 288 MW; 2,522,880 MWh/year; A$201,830,400/year. These are constant-load energy-only scenarios, not forecasts or full electricity bills.
Connection constraint
Headroom is 8 MW, 0 MW and −38 MW respectively. The full campus exceeds 250 MW by 38 MW. Under the same multiplier, maximum IT load is 250/1.2 ≈ 208.3 MW. The expansion phase has no spare margin under this simplified normal-operation scenario; that is not evidence of resilience under a contingency.
Decision and sensitivity
If full-campus PUE were 1.3 instead, demand would be 312 MW and shortfall 62 MW. A recommendation must therefore depend on tested coincident demand, connection timing and service commitments. A battery can cover a time-limited deficit but requires recharging and sufficient duration; it is not an indefinite source of extra grid capacity.
Review your response
- Quantities: all three phases reconcile in MW, MWh and A$ with declared assumptions.
- Feasibility: connection shortfall and the IT ceiling are calculated correctly; normal-operation headroom is not described as N−1 security.
- Integration: alternatives connect demand, network timing, procurement, cooling and service reliability.
- Decision: the recommendation names evidence gaps and a decision gate rather than claiming a final engineering design.
This checklist supports self-review. No tutor marking or independent credential is included.
Related topics 8
Sources & evidence
AEMO QED for industry context; the separate synthetic worksheet for calculations. A real hourly load profile, connection terms and resilience studies are additional evidence to obtain, not supplied facts.
AEMO digital demand surge