Every few months a new projection lands on the desks of planners in Dhaka, and every one of them points the same way. Bangladesh uses more electricity each year than the year before, and the gap between the power the country can reliably generate and the power it will need is widening faster than new plants can be built. The question is not whether the country needs more capacity. It is where that capacity can come from quickly, cheaply and without deepening the dependence on imported fuel.
This article sets out the numbers, explains why the headline capacity figure flatters the real position, and makes the case that a meaningful share of the answer is already sitting on top of the country's factories and warehouses. It also sets out, honestly, what rooftop solar cannot do.
The numbers behind the headline
Bangladesh has roughly 25,500–26,700 MW of installed generation capacity. Peak demand is projected to reach 30,000–33,000 MW by 2030 and around 52,000 MW by 2041. Renewables currently supply only about 5–6% of the electricity mix.
Put those figures side by side and the scale of the task becomes clear. Even if every megawatt installed today stayed online and ran whenever it was needed, the country would need several thousand megawatts of new capacity before the end of this decade, and something close to 25,000 MW more by 2041. That is roughly the size of the entire current fleet, added again, in less than two decades.
Installed capacity vs projected peak demand, MW
2030 shown as the midpoint of the 30,000–33,000 MW projection.
Why installed capacity overstates available power
Installed capacity is a nameplate number. It tells you what the fleet could produce if every unit were available, fully fuelled and running at its rated output at the same moment. That never happens anywhere in the world, and in Bangladesh the difference between nameplate and dependable output has been visible in every summer of load-shedding.
Three factors do most of the damage. First, a large part of the fleet runs on gas, liquid fuel or imported coal, so its output depends on fuel supply and foreign-exchange availability rather than on engineering. Second, older units are derated: a plant built for a given output may deliver noticeably less after decades of service. Third, planned and forced outages mean a slice of the fleet is always offline for maintenance or repair.
The practical effect is that the headline gap between capacity and demand is the optimistic version of the problem. The gap that factory owners actually feel, measured in hours of outage and litres of diesel, is larger.
There is a cost dimension too. When fuel has to be imported and paid for in foreign currency, every unit generated from it carries exchange-rate risk. That risk eventually reaches consumers through tariff revisions, and it reaches manufacturers first, because industrial and commercial tariffs are where cost recovery is most visible. For an export-oriented RMG plant competing on price per garment, a grid tariff that can move with the dollar is a strategic exposure, not just an operating cost.
“The gap a factory feels is not measured in megawatts. It is measured in stopped lines, spoiled batches and diesel receipts.”
Where rooftop solar fits
No single technology closes a gap of this size. But rooftop solar on commercial and industrial buildings has three characteristics that make it unusually well suited to the part of the problem that has to be solved first.
It is fast
A new utility-scale thermal plant typically takes five to seven years from concept to commercial operation, once land, financing, fuel contracts and grid connection are in place. A well-prepared 1 MWp factory rooftop can move from site survey to first generation in around 16 weeks. Speed matters because the demand curve is not waiting for the planning cycle, and because every year of delay is another year of peak-season load-shedding for the plants that can least afford it.
It sits where the load is
Industrial clusters around Gazipur, Narayanganj, Savar, Ashulia and Chattogram are where much of the country's daytime demand is concentrated. Generation on the roof of the building that consumes it avoids transmission losses, eases pressure on congested distribution feeders and needs no new land, which in one of the most densely populated countries on earth is not a minor point.
It is cheap at the point of use
Once built, a rooftop plant produces electricity with no fuel cost. Over a 25-year design life the levelised cost of that energy is well below the blended commercial tariff most factories pay, and far below the Tk 28–32 per unit that diesel backup typically costs. The owner also locks in that cost, which makes it immune to future tariff revisions.
Sizing the opportunity
How much of the gap could rooftops realistically close? Nobody has a precise national inventory of structurally suitable industrial roof space, so any estimate should be treated as a scenario rather than a forecast. The table below shows what different levels of adoption would mean using the rule of thumb above.
| Scenario | Suitable roof area | Capacity | Annual generation |
|---|---|---|---|
| Early adopters | 20 million sq ft | ≈ 220 MWp | ≈ 290 GWh |
| Mainstream RMG & industry | 100 million sq ft | ≈ 1,100 MWp | ≈ 1,440 GWh |
| Broad C&I adoption | 250 million sq ft | ≈ 2,800 MWp | ≈ 3,600 GWh |
Even the most ambitious scenario does not close a 25,000 MW gap on its own, and it should not be presented as if it could. What it does is shave the daytime peak in precisely the industrial areas where outages hurt output most, while the country builds the longer-lead generation, storage and transmission it also needs.
What rooftops can't do
Rooftop solar has hard limits, and a serious energy strategy has to plan around them rather than ignore them.
- It does not generate at night. The evening peak, when residential demand climbs, is not served by solar without storage.
- Output falls in the monsoon. Generation in July and August is materially lower than in March and April, so annual yield figures need to be read as averages.
- Not every roof qualifies. Lightweight or corroded sheet roofs, poorly spaced purlins and heavy shading rule out a significant share of buildings unless the structure is reinforced.
- Grids need to be ready. High shares of distributed solar on a single feeder require protection settings, export control and utility coordination.
These limits are why we treat rooftop solar as one layer in a stack. Battery storage moves daytime generation into the evening and rides through load-shedding. Floating solar uses water surfaces where land is scarce. Utility-scale IPP projects add bulk generation to the grid. The combination is more resilient than any one of them.
What changes when thousands of roofs go solar
The benefit of rooftop solar is not only private. When a factory generates a large share of its own daytime electricity, it stops drawing that power from a feeder that is often shared with other industrial users. Multiply that across an industrial cluster and the local distribution network gains headroom at precisely the hours when it is most stressed.
Every unit generated on a roof is also a unit that does not need imported fuel. At national scale, a few gigawatts of distributed solar reduces the volume of gas, oil and coal the country has to buy abroad, which eases pressure on foreign-exchange reserves. And because generation is spread across thousands of sites rather than concentrated in a handful of plants, the system as a whole becomes less vulnerable to any single outage.
There is a commercial dimension for exporters as well. International apparel buyers increasingly ask suppliers for documented renewable electricity and verified emissions data. A monitored rooftop plant produces exactly that record, month by month, in a format auditors can check.
What this means for a factory owner
For an individual plant, the national picture translates into a simple proposition: the grid is unlikely to become more reliable or cheaper in the near term, so the cheapest and most dependable unit of electricity you can buy may be the one generated on your own roof. The decision is worth approaching in a disciplined sequence.
- Collect twelve to thirty-six months of bills so your load profile and tariff history can be analysed properly.
- Commission a structural roof survey before anyone quotes a system size. Purlin capacity, sheet condition and wind uplift decide what is buildable.
- Decide what you are solving for: lower bills, backup through load-shedding, a buyer's renewable target, or all three. That choice determines on-grid, hybrid or storage-led design.
- Compare financing models, including capex purchase, RESCO lease and PPA, on cash flow rather than on headline price.
- Insist on monitoring and O&M from day one. A plant that is not watched quietly underperforms its forecast.
The bottom line
Bangladesh's 30,000 MW question will be answered by many technologies over many years. But the first few thousand megawatts of the answer do not need new land, new fuel imports or a seven-year construction programme. They need structurally sound roofs, careful engineering and owners who decide that their electricity bill is a cost they can control. That work can start this quarter.
Written by
Markets & Policy Desk
Business development, Emerged Power
Tracking tariffs, net-metering rules and power-sector planning so factory owners don't have to.