India’s Renewables Target Falls Short of Growing Demand: Planning and Time-of-Day Balancing Gaps

Executive Summary

Can India Balance Growth in Electricity Demand Purely from Renewable Energy (RE)?

India aims to achieve 500 gigawatts (GW) of installed electricity generation capacity from non-fossil sources by 2030, necessitating more than a doubling of the non-fossil capacity installed as of late 2024. Most of this growth is planned from renewable energy (RE), specifically solar and wind. This non-fossil capacity target surpasses India’s commitment for 50% of its electric power installed capacity to derive from non-fossil fuel-based energy, as part of its Nationally Determined Contribution (NDC) to the United Nations Framework Convention on Climate Change (UNFCCC). This NDC target is based on capacity, and an important question is: what does it mean in energy terms? In this paper, we translate energy and capacity and incorporate demand projections through 2030 and 2036 to examine the sufficiency of RE to meet demand growth. We not only study the energy balance but also consider the time-of-day (ToD) matching between incremental demand and incremental RE output, which will be critical for grid stability and energy security. If RE growth is insufficient to match demand growth, alternative sources will be required, with coal remaining the backstop for India, even after accounting for modest growth in nuclear power and hydro.

This study examines the (mis)match between any new intermittent, also known as variable renewable energy (VRE), generation and demand growth profiles, scrutinising annual, seasonal, daily, and hourly surpluses and deficits. We use 2023 as a base year for annual energy requirements and historical data for RE supply and demand patterns.

There are two dimensions to both supply and demand that we focus on (Figure ES-1). First, there is the aggregate quantum of energy, measured in units of electricity or kilowatt-hours (kWh) Second, there are profiles that vary over times of day and seasonally, which are required for balancing the grid (balancing supply and demand) at all time periods. This imposes stricter requirements (see the right side of Figure ES-1) than simply having “enough energy” measured by the annual average kWh.

Another contribution of this study is the identification of issues of uncertainty and their relative importance. ToD and variability are key issues on both the demand and supply sides. Our study finds that the assumption of demand profiles (load shapes) and their evolution through 2030 and 2036 is a key variable. We model demand growth across a range from pro-rata growth of demand (matching 2022’s shape) to the load profile (ToD shape) shift observed between 2019 and 2022, where more growth was relatively solar-aligned, with higher midday growth than average. Other key variables include (1) the ratio of installed capacity of wind and solar; (2) RE output, measured by the PLF or capacity utilisation factor (CUF); and (3) the shape of RE output for a given capacity. Although this is a national analysis, we also utilise forward-looking VRE output profiles that anticipate technological improvements, such as higher hub heights for wind and more Direct Current to Alternating Current (DC-AC) oversizing for solar, rather than relying solely on historical RE generation data. All these variables lead to different scenarios that we model.

Matching Demand Growth Requires High RE Deployment—India’s RE Targets are Insufficient

We begin with an annual energy-basis adequacy assessment: can the expected growth in electricity demand be met by the planned growth of wind and solar, or a different mix of wind and solar? Stated another way, how much wind and solar capacity is required to meet incremental demand through 2030 and through 2036 (purely on an annual energy basis)? To answer this question, we must assume power demand growth, treated as an exogenous variable, where we take 6.16% and 5.52% as the demand growth rates for 2023–2030 and 2030–2036, respectively, as the compound annual growth rate (CAGR) for the two time periods. This aggregate energy-basis balancing is distinct from the hourly or ToD basis balancing considered subsequently, where both supply and demand shapes are varied under different scenarios.

Figure ES-2 shows the required new RE capacities (solar and wind) to meet new demand in 2030, presented in a waterfall diagram. The base scenario uses demand projections from the 20th Electric Power Survey (EPS) by the Central Electricity Authority (CEA) and RE output based on historical data for 2023, with RE growth at a 2:1 capacity ratio for solar to wind. We then vary a range of supply and demand variables, some of which increase the required RE addition, while others decrease it. The top of the graph in Figure ES-2 shows a line indicating the corresponding total non-fossil capacity, including hydro, nuclear, and other RE sources like biomass power. For the base scenario, we use existing capacities as of
December 31, 2023.

If future CUFs remain the same as historical levels, and no additional capacities beyond wind and solar are added, then, working backwards, 436 GW of solar + wind (“new RE”) would be required to meet demand growth as per EPS, corresponding to 624 GW of non-fossil capacity by 2030. Among the variables that can change, high CUF RE helps lower capacity requirements the most. However, EPS’s demand projections for 2030 appear low, and more realistic growth rates based on the actual 2023 demand as a baseline increase the RE requirements. The growth of rooftop solar (RTS), expected to accelerate under the PM Surya Ghar Yojana, raises the required RE capacity since RTS has a much lower CUF than historical solar, which has predominantly consisted of large solar farms.

Assuming the changes along the waterfall, even with all nuclear and hydro additions under construction being completed by 2030, 348 GW of solar and wind would need to be added, resulting in a total of 559 GW of non-fossil capacity. This indicates that the total non-fossil capacity falls short by about 11.8%, or the growth target for RE (specifically, wind and solar) falls short by approximately 17%.

Since India’s RE targets (the dominant fraction of the 500 GW non-fossil target) fall short, this implies a greater burden on the existing stock of fossil fuel capacity, or even the need for alternative capacity additions. Additionally, if 25.44 GW of under-construction coal plants become operational by 2030, this reduces the remaining demand growth to be met via new RE, resulting in a total non-fossil capacity of 499 GW for 2030 (considering this growth of coal capacity as a locked-in or sunk cost). On paper, this aligns with targets but (a) still implies more coal capacity; (b) assumes everything proceeds as planned (especially higher CUFs for new RE than historical); and (c) overlooks the challenges of ToD matching.

Converting the total 2030 RE growth requirement into annual capacity addition requirements means over 48 GW of annual RE additions would be needed through 2030, which is more than double the historical achievement thus far. Even this is after accounting for under-construction new coal capacity, as shown. Historical shortfalls in RE additions have led to strains on the coal generation system and the need for more coal power plant capacity as a backstop.

Time-of-Day Balancing is Far More Complex and Challenging

Even if sufficient RE capacity exists to meet demand on an annual energy basis, this remains true only “on average.” It does not account for the variability of RE, which is critical for grid planning—the intermittency of both wind and solar is partly predictable but partly stochastic (random). Spreading RE generation across wide geographic areas helps, but as state-level analysis indicates (in a separate forthcoming study), this is true only up to a point due to inherent correlations in RE output. For solar, the reasons are more apparent, but for wind, many parts of India share the same underlying driver (the monsoon).

Using wind and solar output data available at state levels and for new high CUF RE sites, we compare the RE output with demand on an hourly or 15-minute time-block basis. Given uncertainty in the demand shape (also known as the load profile) over time, we examine multiple scenarios that span a proportional or pro rata demand growth shape based on our base year, as well as different shapes according to different years of growth. Inherently, a VRE-only system will, by definition, experience periods of both deficit and surplus supply. An energy-basis sizing of RE capacity begins with the total annual deficit equalling the total annual surplus. However, our findings indicate that the levels of the surplus and deficit vary according to the solar-to-wind ratio and the evolution of the load curve. We also find there is a sweet spot of higher wind than current plans (which are the national renewable purchase obligation, or RPO) that lowers the total surplus and deficit across different time blocks.

A simplistic worldview would suggest that if the surplus and deficit are equal in different time blocks, all we need is a technology like storage to time-shift such energy. We analysed the feasibility of this independent of the costs of doing so. The first implication of storage is the need for more RE to overcome storage efficiency losses. Any storage system has round-trip efficiency losses varying from about 10% for a battery to about 25% for hydro pumped storage projects (PSPs). Even assuming RE supply is sufficient to overcome this efficiency penalty, we find that seasonality is a strong limitation for relying on storage.

Given that most upcoming battery systems and PSPs are designed for only a few hours of utilisation each day, and surpluses typically occur in the middle of the day when solar generation is high, we examined the daily surplus and deficit to assess the feasibility of common storage systems. These systems generally have economics and planning based on almost full daily utilisation. Unfortunately, we observe a significant mismatch between days with many hours of deficit and days with many hours of surplus (Figure ES-3). The seasonality is evident on the left, and this becomes even more pronounced when we reorder (stack) the deficit along a load duration curve (LDC), alongside the corresponding surpluses (right graph).

This daily mismatch is only partially mitigated through an optimal wind–solar ratio or a favourable demand shape. There are still many dozens of days where an “average”-sized battery is insufficient. This finding has significant implications for upcoming storage bids in India, many of which are solely for capacity (hardware), leaving the responsibility of charging the battery to the State or distribution company (discom). “Surplus RE” cannot charge the battery on all days and will necessitate either additional RE—which will be costly as it is not required in full every day—or the use of more fossil fuel overnight to charge the battery. However, even if one were willing to accept higher, albeit hopefully occasional, fossil fuel use to charge a battery, on the most critical days, there will be no surplus generation capacity from fossil fuels either. This does not imply that storage has no value in the grid; rather, it will be inadequate for grid balancing at the volumes currently being planned (several hours of storage for the battery sizing). Another separate study we are conducting (forthcoming) focuses on the possibilities and limits of oversizing RE to reduce periods of deficit supply. We know this has value because, while it increases costs, the higher cost is only due to any surplus generation that would need to be curtailed (discarded) in some time blocks, which might be in the order of Rs 3/kWh. In contrast, the value of avoiding a deficit time block is much higher (we are avoiding expensive diesel or a blackout, ostensibly Rs 20+/kWh, or storage, which is also costly).

Policy Implications and Recommendations—More RE but Focus on ToD

First and foremost, India must aggressively increase its deployment of wind and solar energy. It is behind not only its 2030 RE targets but also the RE levels of deployment that can be integrated into the grid even without requiring any storage or reaching the full limits of flexible (part-load) coal operations during the middle of the day when solar generation is high. It also cannot lose sight of the need for planned and even higher growth of nuclear power and reservoir hydropower. The latter is especially important in a high-RE future because of its inherent storage capacity and flexibility.

In the short to medium term, before India considers winding down existing fossil fuel generation, even to plateau fossil fuel growth, it will need more RE capacity than a simple annual energy balancing target would suggest. This is not only to create a buffer or oversize capacity to reduce periods of deficit but also because any use of storage technologies inherently results in significant losses, as does long-distance transmission. This is also before accounting for any RE required for green hydrogen, which could demand over a third of the planned growth in RE capacity for the grid. Not all the energy for green hydrogen would be additional captive supply by the developer.

This analysis demonstrates that India’s 2030 RE targets, even if achieved, will likely fall short of meeting projected electricity demand growth, particularly when considering the critical challenges of ToD balancing and storage limitations. A more nuanced but aggressive approach to RE deployment and grid management is essential.

There are also other nuances that require explicit planning, including those related to location, transmission, solar-to-wind ratio, and the rise of RTS (with its lower output). Most importantly, supply must meet demand in real time, and India needs to gain a better understanding of projected demand profiles in the future. Experts across civil society and academia can contribute, but they require high-resolution (temporal and spatial) data, which is difficult to obtain. For instance, we do not yet have a clear understanding of the true PLF of RTS across diverse locations on a monthly basis—ideally, we need 15-minute data.

There are many additional questions, such as how to plan for and manage behind-the-meter generation like RTS, captive power, open access, agricultural solar, cooling, Electric Vehicles (EVs), and green hydrogen. It is time for not just more modelling but better and more transparent modelling. These factors will also shift based on policies, which are evolving. As we begin to scale storage, we must explicitly plan for the question: “How will it get charged?” It is a myth that “surplus RE” can charge our batteries on all days. Another finding from our work highlights the need to plan for wider variability across years—some years are worse than others, approximately following a seven-year cycle (based on climatic patterns). Thus, many actions, such as two-hour storage, are merely a starting point.

The most important change in planning must be an emphasis on ToD considerations. We need to move away from the levelised cost of energy (LCOE) as a marker of the best supply options and instead factor in system-level costs. Pricing must shift from averaged or socialised costs towards more granular costs, including ToD pricing, for both wholesale procurement and consumer retail supply. These measures will incentivise not only storage but also demand response programmes and other demand-side measures.