09 · Risk
What actually kills this deal
The original risk register lists geothermal dry holes, grid delay, water and FX. That list is correct and incomplete. The fatal risks are offtake in a hydro-long system, first-of-kind CSP delivery, and a model that cannot service 60% debt on honest megawatt-hours.
Risk register — recast
| Risk | Likelihood | Impact | Mitigation that is real |
|---|---|---|---|
| EEU cash gap (FY25/26) | Certain | Deal-breaker | Br 14.69bn shortfall. Do not repay debt from retail margin. MoF wrap + USD capacity. |
| PPA reopen (Corbetti, Tulu Moye) | High | High | Both rewriting tariff as of 15 Sep 2026. Grant the first wells, or a stranded-well clause. |
| Dry-year curtailment | High | High | Sep 2026: inflows −20%, miners cut to 23%. Sell dry-season capacity. Refuse miner offtake. |
| EEP offtake / convertibility | High | Deal-breaker | MoF guarantee, 3-month LC, USD waterfall, political-risk insurance (MIGA/ATI). |
| GERD surplus compresses energy value | High | High | Sell capacity + dry-season shaping, not bulk MWh. Or wait for export contracts. |
| CSP EPC cost / delay | High | High | Do not bid 200 MW. Option the CSP after a measured DNI year and a Chinese/Spanish EPC shortlist. |
| Tendaho dry or cooler wells | Medium-high | High | Phase 1: 3–6 wells, GRMF drilling grant, stage-gate before power block. |
| Water / fluoride / brine | Medium | High | Pump tests before greenhouse FID. Dry-cool CSP. Independent ESIA. |
| Afar security & access | Medium | High | Logistics via Djibouti. Community agreement with Afar pastoralists is not optional CSR. |
| FX (USD/ETB ~163) | Certain | High | 100% USD capacity payment; ETB only for local O&M with indexation. |
| 150 ha too small for 200 MW trough | High | Medium | Re-survey. Troughs need ~2–2.5 ha/MW plus TES, roads, setbacks. |
| Carbon revenue | High | Low $ | Do not gear the debt on Verra. Hydro grid = weak additionality. |
| PwC “validation” reliance | Certain | Reputational | Commission a real model audit (ISAE 3402 / actual audit firm engagement letter). |
Failure modes we actually ran
- Geo CF → 0 (dry field): independent 250 MW still has CSP+PV, but DSCR collapses further; Phase 1 is then a 50 MW PV plant and should be underwritten as one.
- Tariff $0.07: full plant is not a project. Only grant-heavy PV survives.
- CAPEX +15% on CSP: classic African first-plant outcome. Kill 60% leverage.
- 12-month delay: IDC on $1.1B debt at 7.5% is ~$85M. The original “−2.9% IRR” sensitivity is too kind.
- No grants: 10% ($186M) is an aspiration (GCF/AfDB/GEF). Model 0% as downside, not 10% as base.
Legal rails that do exist
PPP Proclamation 1076/2018, Investment Proclamation, Geothermal Resource Development Regulation 453/2019, Electricity Tariff Setting Methodology, and MoF Directive 1067/2025 on power-sector cost transparency. These are usable. They are not a PPA, not a generation license, not a landholding certificate, and not an ESIA approval — all of which the due-diligence memo in the source pack asserts as if they were already in hand.