Paradox #31
The Discount with a Sunset Clause
→ Dorjilung is being financed at roughly half the global average per-kW cost. Either Bhutan's geology is extraordinarily favourable — or the project will experience significant cost overruns later.
Referenced as sidebar in Chapter Two
World Bank loan to Bhutan for Dorjilung hydropower project (May 2026)
USD 515M for 1,125 MW = ~USD 458/kW
Global average hydropower capex per kW
USD 900-1,200/kW
The full numbers
The World Bank loan of USD 515M was signed in early May 2026 for the Dorjilung hydropower project (1,125 MW). On a per-installed-kW basis, that’s about USD 458/kW. Global hydropower capex benchmarks: large projects in mountainous terrain typically cost USD 900-1,500/kW; mid-sized projects USD 800-1,200/kW. Bhutan’s PHPA-I, by comparison, has experienced major cost overruns — original estimate Nu 35B, current estimate 3-4x that (paradox #3).
Imagine this
The Dorjilung project breaks ground in 2027. The contractor’s bid was based on the USD 515M envelope. Three years in, geological surprises emerge. Five years in, the contractor requests cost variation orders. Seven years in, the project is 60% complete and the original USD 515M is fully spent. The choice: (a) abandon, (b) negotiate additional financing, (c) bring in supplementary contractors. By 2034, the project may have cost USD 900M-1.2B — closer to global benchmarks. The additional USD 400-700M comes from somewhere: more bilateral debt, multilateral lending, sovereign guarantees, equity infusion. Whichever source, the country’s debt stock grows beyond what was committed at signing.
Where this came from
Hydropower projects globally are notorious for cost overruns — the literature documents a ~30% average overrun rate on large dams, with some projects (Three Gorges, Belo Monte) experiencing 50-100% overruns. Bhutan’s PHPA-I track record suggests overruns may be worse here than global average.
The Dorjilung USD 458/kW figure, if it holds, would be exceptional. If it doesn’t hold (the more likely outcome), the eventual cost will be borne by Bhutanese debt service.
Why this matters now
The 13th FYP commits to 10 large hydropower projects beyond PHPA-I/II. If each experiences PHPA-style cost overruns, the eventual debt burden is substantially higher than current FYP projections (which assume ~Dorjilung-style benchmark capex). The USD 26B planned through 2040 could become USD 35-50B in actual deployment — pushing public debt well above the 94.8% of GDP FYP projection.
What it should be
Project budgets should incorporate realistic overrun expectations. A 30-50% contingency on USD 515M would be ~USD 670-770M as a more honest budget. International contractors should be selected based on track record at Bhutan-specific conditions, not lowest bid.
How others do it
- China’s Three Gorges — original estimate USD 17B; final cost ~USD 30B (76% overrun). Built faster than expected (15 years) but cost-wise overran significantly.
- India’s Sardar Sarovar — projected costs grew 4-5x over the project lifetime due to delays and design changes.
- Norway — exceptional track record of on-budget hydropower delivery, partly because Norwegian engineering firms specialise in cold-climate complex projects.
- Bhutan PHPA-I — 3-4x cost overrun, 18+ year duration. Bhutan PHPA-II — 14 years, lesser overrun.
- Dorjilung at USD 458/kW would be among the cheapest hydropower per-kW deployments globally — a flag, not a feature.
The question we should be sitting with
Is Dorjilung’s budget realistic, or is it the cost the financing community wanted to hear? What happens if the real cost is USD 1B+, and what does that mean for the next 10 hydro projects in the pipeline?