The Punatsangchhu Problem

If you are here for the short answers: Punatsangchhu-II is finally running. The 1,020-megawatt plant — PHPA-II, the one Bhutanese shorthand calls phpa2 — was commissioned in December 2024, seven years past its original 2017 target, and in 2025 delivered its first full year of generation, about 3,676 million units. In May 2026 its power-purchase agreement with India was signed: exports at Nu 5.10 a unit, stepping up on a fixed schedule to a levelised Nu 5.67 over the 35-year contract — roughly double what the older dams earn. Against a project cost that ended near Nu 94 billion, it is, at last, earning.

And Punatsangchhu-I is still not. The bigger sister — 1,200 MW, groundbreak 2008 — remains uncommissioned eighteen years on, its budget up from a projected Nu 35 billion to roughly Nu 100 billion and counting.

Those are the facts a search engine can give you. The more useful question is what the two dams, taken together, say about the machine Bhutan builds its future with — because the country has just committed to running that machine ten more times.

One river, two lessons

The Punatsangchhu was chosen for good reasons: a big, steep, reliable river in the west, close to the Indian border and the Indian grid. In the late 2000s Bhutan and India agreed to develop it with twin projects — I and II, a few kilometres apart — under the model that built the country: India finances (a mix of grant and loan), Indian firms construct, Bhutan repays in electricity, and the tariff is set cost-plus, in rupees.

PHPA-I became the geology lesson. In 2013 the right bank above the dam pit began to move — a slow toe-slide in rock the surveys had read as sound. Shear zones kept appearing; slope after slope was re-engineered; the active mountain-building that raised the Himalaya turned out to be still at work directly beneath the dam. Commissioning slipped past 2016, past 2020, past every revised promise since. Eighteen years is now twice the time China needed to build a dam thirteen times larger. The engineers who joined at groundbreak as new graduates are in their mid-forties.

PHPA-II became the accounting lesson. It escaped the worst geology and got built — fourteen years, roughly double a global-pace schedule, but built. Then the Royal Audit Authority worked through the books. Its FY 2024–25 report flags Nu 7,790 million in irregularities at PHPA-II alone — 78% of everything flagged in the entire country that year, more than the rest of the public sector combined: steel-price escalation indices applied wrongly (over Nu 125 million in overpayments in a single year), unreconciled contractor variations, finance-management failures accumulated across seventeen years and four prime contractors. Not necessarily theft — mega-projects everywhere produce findings at this scale when auditors finally get through — but that is rather the point: the model produces this even when it works.

What the tariff really settles

The PHPA-II tariff deserves a closer look than it usually gets, because it quietly answers the question every Bhutanese asks about the dams: was it worth it?

The answer the contract gives is: narrowly, in rupees, over 35 years. Nu 5.10 rising to a levelised 5.67 is a cost-plus number — project cost, financing, O&M, depreciation — benchmarked to Indian norms. It is double the old Chhukha-era tariffs, which reflects what the dams now cost to build. But two clauses that are not in it matter more than the ones that are. There is no dollar clause: the revenue is rupee-denominated for thirty-five years, so its hard-currency value inherits the rupee’s long slide — the same erosion that has already cost the older plants most of their real earnings. And the escalation steps — 7%, then 5%, every five years — sit below recent Indian inflation, meaning the tariff is designed to fall gently in real terms across its life.

In other words: PHPA-II will pay its debts and fund the budget, in a currency built to soften. It is a good deal by the standards of the model. Whether the model’s standards are good enough is the accidental currency bet, asked with concrete and rebar.

The pipeline bet

Here is why the Punatsangchhu ledger matters beyond one river. The 13th Five-Year Plan commits Bhutan to ten further large projects — 9,892 MW, from Kholongchhu (600 MW) through Dorjilung (1,125 MW) to the giant Sankosh (2,585 MW) — on the order of Nu 800 billion of construction over the coming decades, mostly under variants of the same arrangement: partner financing, rupee tariffs, cost-plus pricing, Himalayan geology.

Run the two Punatsangchhu outcomes forward. If the next ten build at PHPA-II pace (fourteen years, heavy audit tail), the 2040 generation targets land closer to the 2050s, with a compliance workload the six-auditor teams of today cannot absorb. If any build at PHPA-I pace, engineers now in college will retire before the turbines turn. The optimistic case — that the lessons are learned, the surveys go deeper, the contracts get tighter, the escalation formulas get checked in real time — is genuinely possible. But it has to be built, institutionally, and the evidence that it has been is thin: same financing partners, same construction firms, often the same teams.

None of this is an argument against the river. Bhutan’s water remains its one world-class resource, and the economy runs on it — which is precisely why the machinery that converts water into money deserves harder scrutiny than the celebration of each commissioning allows. PHPA-II’s first exported unit took sixteen years from approval to arrive. The question the Punatsangchhu has spent two decades asking is not whether Bhutan can build dams. It is whether Bhutan can afford how it builds them — ten more times, on a warming river, in a melting currency.