Q1 business case·BCDF Stage 2 — Preliminary Business Case·Gateway Gate 1·2026
BCDF Stage 2 takes the substantiated need from the SBC and runs a structured options analysis — longlist to shortlist to a single reference project. Four decision axes were screened against real alternatives. Every axis selected the configuration now in the deal room; the preferred option is a behind-the-meter pilot at Swanbank, staged to the 800 MW program. Recommendation: proceed to a Detailed Business Case.
Options were screened on the binding constraint for sovereign compute — firmed power + grid headroom + a heat sink under one cooperative owner — not on rack availability. The longlist (do-nothing, colocation, greenfield, brownfield energy-precinct co-location) reduced to a shortlist on siting, then cooling, power, density and staging.
| Option | For | Against | Verdict |
|---|---|---|---|
| Do nothing | No capital at risk | Need unmet; QGOV defers AI or accepts offshore exposure | Rejected |
| Colocation (NEXTDC-class) | Fast to first rack; no construction risk | Rents someone else's power story — $/kW premium, no behind-the-meter energy, no heat-sink control, sovereignty diluted by the landlord's stack; caps below program scale | Rejected |
| Greenfield build | Full design control | Grid-connection queue in years; no existing firming; land, water, approvals from zero | Rejected |
| Swanbank brownfield | 385 MW firm generation + 250 MW/500 MWh battery operational since Feb 2026; 1.2 GW connection with a commissioned 275 kV substation; purpose-built cooling reservoir; one QLD-government-owned landlord courting tenants; 11 km from Brisbane | Brownfield legacy (PFAS) to scope; terms not yet executed | Preferred |
This selection is why the prior NEXTDC-colocation Series A page was superseded.
| Axis | Alternatives screened | Selected | Why |
|---|---|---|---|
| Cooling | Air (fails at ~80 kW) · evaporative (potable draw, discharge, PFAS interaction) | Closed-loop direct-to-chip liquid to the Cooling Water Dam | Zero potable water per token; no evaporative draw, no creek discharge; PFAS contained by construction |
| Power | Grid retail supply (wholesale + network exposure) | Behind-the-meter firmed block, battery-first, gas transitional | Below the ~$65/MWh QLD grid average; electricity <3% of revenue |
| Density | 20–40 kW air-cooled racks | ~80 kW liquid-cooled racks | Halves hall footprint per MW; matches the loop |
| Staging | Build big first (100 MW+ day one) | Gate-built staircase: ~2 MW → ~20 MW → ~100 MW → 800 MW | Each step built only when the demand stack underwrites it; capital never runs ahead of proven utilisation |
A ~2 MW, ~2,850-GPU behind-the-meter pilot at CleanCo's Swanbank Energy Precinct, closed-loop liquid cooled to the Cooling Water Dam, firmed battery-first — staged through 20 MW and 100 MW phases toward the 800 MW program on the same 1.2 GW connection. This is the reference project the DBC develops.
| Installed GPUs (~25 racks @ ~80 kW, ~2 MW IT) | ~2,850 |
| Annual revenue (~$2.35/GPU-hr · 85% utilisation) | ~$50M |
| EBITDA (~83% margin) | ~$41M |
| Build capex | ~$98M |
| All-in Stage-1 raise (all-equity worst case) | ~$116M |
| Simple payback · electricity share of revenue | ~2.4 yr · <3% |
Stress-screened to $1.80/GPU-hr · 65% utilisation — remains EBITDA-positive, payback toward ~5 yr. Full grid in the financial model.
| Risk | Class | Treatment carried to DBC |
|---|---|---|
| Offtake not yet contracted | Commercial | Tenancies over firmed-MW blocks; QWork Tenancy 01 anchor; executed tenancy an investment-decision condition precedent |
| Site terms TBC (lease, $/MWh, dam, connection) | Commercial | Wave-1 workstream; all executed before the investment decision |
| PFAS legacy contamination | Environmental | Closed-loop contains by construction; scope signed with CleanCo's environmental team |
| Gas vs sovereign-clean narrative | Reputational | Battery-first firming, gas transitional; auditable renewable path; no 100%-renewables claim |
| GPU price & availability | Supply | Supply agreement with delivery deadline, ramp pricing, pro-rata shortfall relief; vendor finance on the resaleable asset |
| Technology refresh | Technology | 4–5 yr depreciation; resale backstop; tight per-tranche generation |
| Single-site concentration | Operational | N-1 connection; program-scale redundancy addressed at Phase-1 |
| Gate | Gateway review | What it tests | Our product | Status |
|---|---|---|---|---|
| 0 | Strategic assessment | Service need substantiated — sovereign compute for QGOV & the regulated sector | SBC | Addressed |
| 1 | Preliminary evaluation | Options analysed; preferred option (Swanbank, closed-loop, staged) selected | PBC | Addressed |
| 2 | Business case | Detailed evidence — economic, financial, commercial, risk, management | DBC | In progress |
| 3 | Contract award / investment decision | Executed terms + tenancy; Stage-1 capital committed (FID) | DBC | Pending — CPs open |
| 4 | Readiness for service | Pilot energised; switch cut over to Swanbank silicon; dashboard → production | Wave 2 | Future |
| 5 | Benefits realisation | Utilisation proven; Customer #2 live; QGOV value delivered | Wave 3–4 | Future |