On 20 May, Bangladesh generated more electricity than ever before: 17,200 MW at 9pm. Even then, around 400 MW of demand went unmet, meaning the country was producing a record amount of electricity and still had load-shedding.
The country’s installed capacity is reportedly 28,494 MW. A system that can produce 28 gigawatts, manages 17 at its best and still cuts power is not short of power stations.
It is short of the fuel to run them. And it is short of the money to pay for that fuel.
A system built in a hurry
When the Awami League government came to power in 2009, power cuts were severe. In October 2010, it passed a law to fast-track new power projects. Projects could be awarded without tenders, and the law also blocked court challenges. The High Court declared parts of it unconstitutional in November 2024.
The results are now clear. A government review of the power sector found that capacity grew more than fivefold between 2009 and FY2024, but the plants were running at only 40 to 50 per cent of their capacity.
The committee put idle capacity at 7,700 to 9,500 MW. Payments to power producers went up elevenfold, while electricity generation rose only fourfold. Capacity payments increased almost twentyfold.
The basic argument for that rapid expansion was simple: Bangladesh needed to end severe power shortages quickly.
But the master plans made things worse. The 2010 plan, prepared with Japanese technical help, aimed for coal to make up 50 per cent of the power mix and gas 25 per cent by 2030.
By 2016, Japan’s aid agency, JICA, had reviewed the plan. It found that actual demand in 2015 was about 8,000 MW, while the forecast had been around 10,000 MW. Gas and coal generation had reached only 40 and 60 per cent, respectively, of what the plan had expected. No coal plant had even started construction.
The revised plan then forecast peak demand of about 21,300 MW in 2026. The actual record is 17,200 MW.
So the plans overestimated demand. They also depended on fuel that did not arrive. That is how a country ends up with idle power plants and blackouts at the same time.
The fuel problem
Gas was supposed to be the foundation, but it has been shrinking.
Domestic gas production was around 2.7 billion cubic feet a day in 2017. It is now below 1.65 billion. Industry experts say no major gas field has been found in three decades.
The energy minister told parliament that gas supply averaged 2,654 million cubic feet a day this fiscal year, up to April. That left a shortfall of about 1,146 million cubic feet.
The response has been slow. A plan to drill 50 wells began in 2022. By late last year, only 20 had been completed. They added 90 million cubic feet a day, against the expected 210 million. Officials said Bangladesh had experience drilling only two or three wells a year. A new rig for the state explorer, Bapex, costing Tk3 billion to Tk3.5 billion, was still pending last year.
The result is a familiar cycle. Domestic gas production falls, exploration moves slowly, and the gap is filled with imports.
Even gas that has already been found is sitting unused. Bhola North, discovered in 2018, had still not started production by mid-2026. Bhola’s wells can supply about 182 million cubic feet a day, but around 47 million cubic feet goes unused because there is no infrastructure to bring it out. A feasibility study for a Bhola–Barishal pipeline has been completed by GTCL. What is missing is the pipeline itself.
Meanwhile, money went elsewhere. Petrobangla planned to spend about Tk579 billion on 115 LNG cargoes in FY2025–26. Bonik Barta calculated that this was 51 times the planned spending on exploration.
Imported LNG filled the gap, and it is expensive. Domestic gas costs roughly Tk3 per cubic metre; imported LNG costs around Tk55. The LNG import bill rose from Tk16,505 crore in FY2020–21 to about Tk59,000 crore in FY2025–26. The Institute for Energy Economics and Financial Analysis (IEEFA) estimates that the share of imports in Bangladesh’s primary energy rose from 47.7 per cent to 62.5 per cent over the same period.
Imports bring their own risks
Imported gas also needs terminals. Bangladesh has two floating LNG terminals off Moheshkhali. JICA warned in 2016 that storage arrangements should allow for bad weather and accidents.
That warning proved relevant. A fire on 21 July shut one terminal and cut about 450 million cubic feet a day from supply. From 26 July to 15 August, load-shedding often crossed 3 GW. On 13 August, loading at the second terminal also stopped because of bad weather. On 24 September, bad weather cut supply by another 300 million cubic feet a day.
Then there was the war. QatarEnergy declared force majeure on 2 March. Oman’s OQ Trading followed on 5 March, and Excelerate Energy on 6 March. Qatar later warned that it might deliver only 20 of the 40 cargoes scheduled for this year. Petrobangla’s finance director said the war added Tk10,600 crore to costs.
The shock came from outside, but it hit a system that already had a gas shortfall of more than 1,300 million cubic feet a day. There was no cushion.
A system built on imports does not cause wars. It imports their effects.
The money problem
Money is where the problem of idle plants meets the problem of blackouts.
The yearly losses of the Bangladesh Power Development Board (BPDB) rose from Tk5,468 crore in FY2015 to Tk50,565 crore in FY2025. It buys bulk power at Tk12.35 a unit and sells it at an average of Tk6.63. Its unpaid bills crossed Tk250 billion early this year.
The effect can be seen in the plants. In April, gas-fired plants with a combined capacity of 12,200 MW produced at most 5,200 MW. Oil-fired plants could not increase output because of unpaid dues. Four major coal plants with a combined capacity of 4,354 MW produced 1,662 MW, or just 38 per cent of their capacity.
IEEFA says the BPDB’s weak finances are pushing it towards load-shedding, even as it continues paying capacity charges to plants that are sitting idle.
In short, the state guarantees payment for capacity. It does not guarantee that fuel will be available.
Relief keeps slipping
Hoped-for relief keeps getting pushed back. The contract for the $12.65 billion Rooppur nuclear power plant was signed with Russia in December 2015. The 2016 plan expected its first unit to start in 2024. Fuel loading began in late April, but faulty safety valves have delayed progress, and the science minister could not give a date. Temporary handover is now planned for December.
Renewables are still small. In April, they stood at 1,059 MW, or just 3.7 per cent of installed capacity. Even if both Rooppur units are running, their combined 2,400 MW would be only about 8 per cent of installed capacity.
The new roadmap is ambitious again. It plans to add 12,940 MW by 2030, including 10,000 MW of solar, for a projected peak demand of 20,000 to 25,000 MW. It also proposes two more floating LNG terminals. The roadmap itself acknowledges that LNG could account for more than 60 per cent of gas use after 2030, and that dependence on imports could rise to about 74 per cent.
Others argue that coal is much cheaper than LNG or liquid fuels and should carry the baseload for now. There are also calls to make better use of Bangladesh’s own gas fields, speed up exploration and expand solar. But the drilling record shows that the domestic route is not necessarily quick either.
The test for the next 12,940 MW
The real failure was not construction. It was what the construction was for. A power plant is not energy; it is a promise of energy. That promise depends on the fuel, the dollars, the grid and the contract behind it.
The current BNP government’s own five-year framework says the $1.5 billion to $1.8 billion annual capacity-payment burden can be reduced through reform.
So the test for the next 12,940 MW is simple. Before a megawatt gets a contract, does it have a fuel supply and a paying buyer?
[Image Credit: IEEFA]
Jannatul Naym Pieal is a Bangladesh-based writer, researcher and journalist with over a decade of experience in professional journalism. He is also the author of 10 published books and a researcher focusing on Bangladesh’s media industry and its intersections with broader social and academic fields.
