Nusa Tenggara’s first solar gigawatt displaces Java’s coal, not local diesel

Summary

Indonesia’s 100 GW solar target is nearly six times the 17.1 GW planned for 2034, so where the first gigawatts land is still an open decision. We test one placement, 1 GW in Nusa Tenggara, where diesel still supplied a quarter of grid electricity in 2024, and ask what it displaces and at what cost.

In our model, Nusa Tenggara’s 610 MW diesel fleet sits idle in 2034, so the gigawatt instead displaces Java’s subcritical coal: 1.81 TWh and 1.78 MtCO2 a year. It cuts the fuel bill by US$56 million at a net system cost of US$28.4 million.

That cost buys four years of acceleration, since the model builds the same solar by 2038 on economics alone. The binding question is where each gigawatt lands and what it meets there, and anyone can test that in Scenario Builder.

Indonesia’s 100 GW solar target is almost six times the 17.1 GW planned for 2034

In August 2025, President Prabowo announced a target of 100 GW of new solar capacity, with 80 GW envisaged as village-scale solar-plus-battery systems for the roughly 10,000 remote communities that remain off-grid, most of them running on diesel generation today — managed by the Merah Putih village cooperatives.

That ambition sits well ahead of formal planning: the current Electricity Supply Business Plan (RUPTL 2025–2034) projects around 17.1 GW of solar by 2034, and the longer-horizon National Electricity General Plan (RUKN 2025–2060) spreads a comparable build-out — 108.7 GW — over 35 years.

Our analysis starts in that gap — and speaks to anyone deciding where the 100 GW should land: what happens if solar moves faster than planned?

Rather than model the full 100 GW at once, we isolated the first gigawatt and placed it in the Nusa Tenggara node — Bali, West Nusa Tenggara, and East Nusa Tenggara. Diesel there is not only an off-grid problem: it still supplied a quarter of the two provinces’ electricity in 2024, and roughly 38% across Maluku and Papua. East Nusa Tenggara also holds Indonesia’s largest provincial solar potential, and the RUKN singles the region out as a priority for solar.

Model and scenario set-up

The analysis uses Scenario Builder’s Indonesia model, which runs into two stages: a capacity-expansion layer decides what gets built each year to 2050, and its 2034 fleet is passed to an hourly dispatch layer that decides what runs. Seven grid regions.

There are four scenarios, built by adjusting the following input parameters:

For full details on the input data, please visit our documentation here.

What one gigawatt changes in 2034

By 2034, Nusa Tenggara’s 0.6 GW of diesel already sits idle

Figure 1 shows where Nusa Tenggara’s power comes from in 2034, with and without the gigawatt. The region’s 22.4 TWh of demand is met by local gas, coal, geothermal, solar, and hydro, plus 3.2 TWh of net imports from Java - while its 0.6 GW of oil-fired capacity generates nothing, out of merit in every hour.

That is not the grid the diesel-replacement framing describes. The 2034 system, built under current plans, has more gas, more coal, geothermal, and a reinforced link to Java - so the gigawatt displaces something else entirely.

The added 1.82 TWh of solar displaces 1.81 TWh of Java’s coal, cutting 1.78 MtCO₂

Figure 2 shows what changes when the gigawatt is added. Two bars move: solar output in Nusa Tenggara rises by 1.82 TWh, and Java’s subcritical coal generation falls by almost exactly that amount, 1.81 TWh. Nothing else in the seven-region system changes, and none of the new solar is curtailed, even without storage.

The gigawatt avoids 1.78 MtCO₂ a year - 0.98 tCO₂/MWh, the emission intensity of subcritical coal, the most carbon-intensive plant on the system. Had the displaced megawatt-hours been gas, the saving would have been less than half: it is set by the marginal fuel.

Nusa Tenggara’s net imports from Java fall 56%, from 3.2 to 1.4 TWh

Figure 3 describes the interconnector carrying traffic both ways. Adding the gigawatt cuts imports and raises exports at once, taking net imports from 3.2 to 1.4 TWh while total traffic on the link rises slightly. The imports that remain fall in the hours solar cannot serve, so the night-time flow from Java is the region’s firm supply; the daytime flow now runs the other way, covering around 2% of Java’s demand.

The gigawatt cuts the fuel bill by US$56 million a year, and costs US$28.4 million net

Coal is cheap to burn, which makes this the harder economic case. The avoided fuel is worth around US$30/MWh; the new solar costs roughly US$46/MWh once capital and fixed operating costs are annualised. Forcing the gigawatt into the 2034 fleet therefore costs about US$28.4 million a year net, around 0.1% of national power system costs, while cutting the fuel bill by US$56 million and emissions by 1.78 MtCO₂. That works out to an abatement cost of US$15.9, or roughly US$16 per tCO₂ avoided.

Left to its own economics, the model builds the same solar anyway as shown in Figure 4: the base scenario crosses 2.2 GW in 2038, four years after the counterfactual, and both pathways converge to around 12.3 GW by 2050 — nationally, 125 GW. The intervention changes when the region reaches each level of solar, not where it ends up.

The US$28.4 million therefore buys roughly four years of earlier fuel savings and emissions cuts as shown in Figure 5. Whether that acceleration justifies the cost is a call for policymakers, not a modelling result; what the modelling adds is the number itself, and it is small.

Three caveats apply to these figures. The model treats Bali and both Nusa Tenggara provinces as one node, so zero curtailment is an upper bound; the results cover one dispatch year, 2034, at default fuel prices; and the gas fleet’s 60% availability limit is what leaves coal, rather than gas, as the marginal plant.

Where each gigawatt lands matters as much as how fast the 100 GW is built

Three findings travel beyond this scenario. The first is that what solar displaces is set by what else the grid can reach. Here, an interconnected gigawatt meets Java’s coal. However on a grid where diesel still supplies a third or more of electricity (as in Maluku and Papua) it would meet diesel — a smaller emissions saving, but a far larger fuel-cost one.

The second is that the economics are a question of timing, not cost. The least-cost plan reaches 125 GW of solar by 2050 on its own; the target moves that build-out earlier, at a price that is measurable and small. The harder question is who carries it - PLN, the budget, or cooperative financing — and that tariff framework, more than the cost of panels, is what separates the target from the plan.

The third is a question the scenario opens but does not settle. The interconnector’s reinforcement is already in RUPTL, so the carbon content of Nusa Tenggara’s night-time imports will be set by Java’s mix: cleaning the sending node counts for as much as building at the receiving one. And Java’s coal is displaced, not retired — whether growing clean imports can eventually support actual retirements is the natural next test.

Test the next gigawatts yourself

This scenario and any variant of it can be reproduced in Scenario Builder. The input data, calibration, and configuration are documented, and moving the same gigawatt to another node, another year, or a genuinely isolated system is an afternoon's work.

Ready to explore Indonesia's solar ambitions? Access the Scenario Builder today to model these alternative pathways yourself.

Start Modelling

Scenario BuilderScenario Builder