A background note can be accessed here: PM Surya Sarovar Yojana: India’s ₹5,070 Crore Push for Floating Solar
The scheme seeks to expand solar generation on reservoirs and industrial ponds, using floating installations to reduce pressure on scarce land. To what extent does avoiding land acquisition make floating solar economically preferable, given the additional costs and uncertainties associated with site preparation, anchoring, environmental assessment and operations on water bodies?
The scheme is directed at the right constraint, and the use of public finance to de-risk the preparatory stage is a particularly welcome feature. Its design, however, can be strengthened at the point of implementation.
Floating installations in India cost about 20 to 30 percent more than ground-mounted plants, with estimates ranging from 10 to 40 percent by site. Tariffs point the same way: eight Indian floating solar awards average ₹3.63 per kilowatt-hour against ₹2.72 for ground-mounted solar, although these span different bid windows and module price cycles. The land saving does not by itself close this gap. Five thousand megawatts of ground-mounted capacity would require roughly 20,000 acres, and at prevailing agricultural values in most states, the cost of this land would be less than the additional cost of floaters, mooring, and anchoring. Lease charges, storm damage liability, and floater replacement sit outside the capital estimate.
We would therefore read land avoidance as a timeline and litigation benefit rather than a cost benefit. That benefit is substantial, since land acquisition is where Indian solar projects most often lose years. Land scarcity is, in any case, a local rather than a national constraint, felt most in states such as Kerala and Bihar. Floating solar becomes competitive where evacuation infrastructure already exists at hydropower and thermal stations, seasonal drawdown is modest, and bathymetry is favourable. Prioritising such sites would do more for project economics than a general land-saving argument.
The financial architecture of the scheme is well constructed. What would complement it is the regulatory scaffolding that determines where projects are located, on whose authority, and on what terms.
Every supported project must include at least two hours of co-located energy storage, linking capacity expansion with the ability to supply electricity beyond periods of solar generation. Does this requirement materially improve the value of floating solar to India's power system, or could a uniform storage mandate impose costs that vary substantially across sites and grid conditions?
The storage requirement is sound in principle, given the changing structure of the grid. The sizing is worth revisiting. The scheme pairs 5,000 megawatts with 10,000 megawatt-hours, or two megawatt-hours per megawatt of solar. The Central Electricity Authority’s advisory of 18 February 2025 recommends two-hour storage equivalent to ten percent of installed solar capacity, or 0.2 megawatt-hours per megawatt. The scheme calls for ten times that, and at current battery prices, the storage would cost roughly twice the total outlay.
A uniform requirement also sits awkwardly at both ends of the site range. On a hydropower reservoir, the turbines already provide firming, and holding water back through the solar day for release at the evening peak shifts energy at very low marginal cost. At an irrigation tank or industrial pond, two hours is modest against an evening ramp closer to four hours. We would suggest that procurement specify outcomes such as peak blocks, ramp rates, availability, and shortfall penalties, and let system studies determine sizing and cycling, with pumped storage and co-scheduled hydropower treated as eligible.
The flat two-hour requirement could be replaced by a delivered peak-energy obligation that is technology-neutral and can be met through batteries, contracted pumped storage, or co-scheduled hydropower at the same reservoir. This avoids paying twice for firming at hydropower sites while avoiding under-provision at standalone sites.
Battery location needs clarification, since siting on water raises fire and thermal considerations, while siting on land reintroduces a land requirement. Where batteries are used, it would help to state that co-location means electrical co-location at the on-land pooling substation, and to specify fire detection, suppression, and containment standards appropriate for a water body.
Reservoir sites are sometimes proposed as natural locations for electrolysis because water is readily available nearby. Water is unlikely to be the binding constraint, since electrolysis consumes about nine litres per kilogram of hydrogen stoichiometrically and 20 to 30 litres at the plant level once purification and cooling are included, which is small relative to irrigation withdrawals. The binding constraints are electrolyser utilisation and proximity to offtake. A floating solar plant on its own operates at roughly 20 percent capacity utilisation, which will not support electrolyser economics. Where a reservoir has a large hydropower station, a firm round-the-clock supply could support high-utilisation electrolysis, but that is a firm-power argument rather than a water argument, and it must be weighed against the value of hydropower flexibility at the evening peak.
The scheme brings reservoirs and other inland water bodies into energy planning, even though these assets already serve irrigation, drinking water, hydropower, fisheries and industrial purposes. How should India determine when a water body is an appropriate site for floating solar and resolve conflicts among its existing uses?
Water falls under Entry 17 of the State List, subject to Union authority over inter-State rivers under Entry 56; electricity falls under the Concurrent List under Entry 38. The Centre can fund projects, but cannot allocate water surface, and allocation frameworks are therefore a precondition.
The 1,200 megawatt project at Jayakwadi in Maharashtra, a notified bird sanctuary with an eco-sensitive zone under the Wildlife Protection Act, 1972, has faced opposition from about 25,000 fishing families and is before the National Green Tribunal. At Omkareshwar, fishermen’s cooperatives have approached the Madhya Pradesh High Court for rehabilitation benefits under the 2013 land acquisition and resettlement law.
A national site-selection and sequencing framework would help settle these questions before projects are announced. Eligible water bodies may be classified by conflict intensity and released in that order: captive industrial and thermal-station ponds, public-sector hydropower reservoirs, and, finally, multipurpose irrigation reservoirs. An accompanying exclusion list would cover notified bird sanctuaries, Ramsar sites, eco-sensitive zones, declared fish sanctuaries, and reservoirs with high inter-annual drawdown variability. Jayakwadi illustrates the value of settling these questions before a project is announced.
Coverage caps should then be defined hydrologically rather than geometrically. The maximum covered fraction is best set with reference to the water spread at the minimum drawdown level, with the developer demonstrating compliance across the full seasonal cycle. Open water should be reserved for wind-driven mixing, gas exchange, and fish movement, and contiguous open corridors are preferable to a single aggregate percentage.
Floating solar should also be explicitly treated under the environmental-clearance framework. It is not separately scheduled under the Environment Impact Assessment Notification, 2006, and environmental and social assessments are therefore largely voluntary. Projects above a defined threshold could be brought into a scheduled category with public consultation, and the resulting commitments could be linked to the disbursement of Central Financial Assistance.
A model water-surface lease and benefit-sharing instrument would complement this framework. States would benefit from a standard instrument covering lease tenure, surface rent, liability for structures during floods, and decommissioning. It could include a defined revenue share for the reservoir-owning department and a ring-fenced livelihood fund for affected fishing communities, with first preference in module cleaning and boat operation contracts, and permission for cage aquaculture in inter-array channels. Compensation for water-surface users currently has no statutory basis, and providing one would allow these questions to be settled at the design stage.
Water benefits should also be measured and verified through a standard protocol. Expectations on water saving also need calibration: field studies indicate 40 to 60 percent evaporation suppression over the covered area, and in reservoirs that fill and spill, the savings are largely converted into spill. The National Institute of Solar Energy could issue a common protocol covering baseline monitoring for at least one full hydrological year before commissioning, and continuing measurement of evaporation, dissolved oxygen, thermal stratification, and fish catch thereafter. A shared method would allow the water benefit to be credited and valued rather than asserted.
Green hydrogen siting should not be driven by water availability. Reservoir sites are sometimes proposed as natural locations for electrolysis because water is readily available nearby. Water is unlikely to be the binding constraint, since electrolysis consumes about nine litres per kilogram of hydrogen stoichiometrically and 20 to 30 litres at the plant level once purification and cooling are included, which is small relative to irrigation withdrawals. The binding constraints are electrolyser utilisation and proximity to offtake. A floating solar plant on its own operates at roughly 20 percent capacity utilisation, which will not support electrolyser economics. Where a reservoir has a large hydropower station, a firm round-the-clock supply could support high-utilisation electrolysis, but that is a firm-power argument rather than a water argument, and it must be weighed against the value of hydropower flexibility at the evening peak.
Sequencing would resolve much of the rest: captive industrial and thermal ponds first, public-sector hydropower reservoirs next, irrigation reservoirs last.