by Amol Phadke, Nikit Abhyankar, and Umed Paliwal.
Despite dramatic declines in clean-energy costs, more than \$1 trillion investments in new coal and gas plants is under consideration worldwide, largely because conventional plants are assumed to be the only economical way to provide reliable, round-the-clock power, such as that required by data centers. India’s recent “thermal-mimic” auction directly tested this assumption by requiring renewables paired with storage to match the reliability and operating profile of conventional plants. Several major developers offered to provide this service at prices below those of conventional power. This calls for reassessing planned coal and gas power investments, especially because renewables combined with storage are also faster to deploy, modular, and cleaner. It also raises a broader question: would consumers, especially large industrial users, be better served by procuring low-cost clean power directly in a more competitive and less regulated market, rather than relying on monopoly utilities to purchase power on their behalf?
Despite dramatic progress in renewable energy and battery storage, more than a trillion dollars of investment in new coal and gas plants remains under consideration globally
The transformation in renewable energy over the past decade has been extraordinary, both in scale and cost. Solar PV and battery costs have fallen by nearly 70-80%, while global deployment has accelerated rapidly. In 2025 alone, the world added more than 600 GW of solar capacity, and solar continues to attract hundreds of billions of dollars of investment annually (IEA, 2026). India is at the forefront of this transformation, adding almost 50 GW of solar in 2025 and overtaking the United States to become the world’s second-largest solar market after China. Solar tariffs in India are now around ₹2.5/kWh, while solar + 4 hours of storage tariffs have reduced to as low as ₹2.9/kWh (SECI, 2025).
There is also growing operational evidence that batteries can support grid reliability at scale. Batteries are already supplying several gigawatts during critical peak periods in Texas, while California now has more than 21 GW of battery resources, increasingly shifting abundant daytime solar generation into the evening peak.
And yet there is an important paradox: power systems around the world are still planning enormous investments in new coal and gas generation. The United States has more than 250 GW of proposed fossil-fuel generation, predominantly gas. China and India are considering more than 200 GW and 100 GW, respectively, of additional coal capacity. Large technology companies such as Meta, despite having net-zero emissions commitments, are choosing to build gas generation capacity at scale. Taken together, these projects could represent well over a trillion dollars of new thermal investments, locking in massive greenhouse emissions for decades.
Why do planners still reach for coal and gas power plants?
The rationale for continued investment in fossil-fuel generation, despite record-low renewable-energy and storage costs, rests largely on two concerns.
First, battery storage today is typically deployed with two to four hours of duration—enough to shift inexpensive midday solar generation into the evening peak. But many major loads, including data centers and industrial facilities, require electricity around the clock. The conventional argument is therefore that storing enough solar energy to supply power through the 10 to 16 hours when solar output is low or zero would be prohibitively expensive.
Second, solar paired with storage is often assumed to be inherently less reliable than coal or gas because of weather variability, including periods of persistent cloud cover. Under this view, renewables and short-duration storage can supply an increasing share of electricity, but the system still requires conventional “firm” resources capable of delivering power whenever needed.
This question is especially important for India. Industry already accounts for a large share of electricity demand and requires substantial round-the-clock supply. Cooling demand increasingly extends well into the evening, while rapid growth in data centers, advanced manufacturing, electric mobility, and other new loads will add further demand for reliable, affordable 24×7 power.
India has already demonstrated that solar plus storage can economically shift cheap midday electricity into the evening peak. But the harder question is different: can renewable energy and storage reliably supply power through the 10–12 hours when solar generation falls to zero—and do so at a cost competitive with new thermal power plants?
How India designed a market test for firm renewable power
That is what makes the Solar Energy Corporation of India’s (SECI) recent 1,000 MW “thermal-mimic” FDRE-RTC auction, among the first of its kind globally at this scale, so important.
Rather than asking developers simply to supply renewable electricity or meet a short evening peak, SECI asked them to bid for a product designed to replicate the operating profile and contractual availability of a conventional thermal power plant. Developers would combine renewable generation and storage to provide firm, dispatchable power under a 25-year contract—similar in scale and duration to a large thermal power purchase agreement (PPA).
The auctioned profile closely follows the way India’s thermal fleet operates today: delivering the most electricity during the evening, night, and early morning, while backing down during solar-rich midday hours.
Generators must supply at least 90% of contracted capacity during six hours nominated by the buyer within the 6 p.m.–10 a.m. window, at least 70% during the remaining non-solar hours, and 50–60% during solar hours. Performance is measured in every 15-minute block, with shortfalls penalized at 1.5 times the contract tariff. This binding 15-minute performance requirement was particularly important because earlier FDRE contracts allowed developers considerably more flexibility in how they met their delivery obligations. The thermal-mimic auction therefore represents a more stringent test of whether renewable energy and storage could reproduce the operating profile of firm conventional generation.
What did the market discover?
The answer was striking: “thermal-mimic” firm renewable power cleared at ₹5.25–5.26/kWh, fixed in nominal terms for 25 years. At this price, renewable energy combined with storage is cheaper than new conventional firm power in India.
This gives India something it did not have before: a competitively discovered market benchmark for renewable power designed to perform much like conventional firm generation. For utilities planning new capacity, the relevant comparison is therefore no longer between intermittent renewables and coal, but between different technologies capable of meeting the same underlying power requirement.
Is the price sustainable?
A natural question is whether ₹5.25/kWh reflects a replicable market price or simply an unusually aggressive outlier bid.
The auction results provide considerable reassurance. Sixteen developers participated, with seven securing capacity and all winning bids falling within the narrow range of ₹5.25–₹5.26/kWh. NTPC Renewable Energy, the renewable arm of India’s largest thermal power generator, bid only 3% above the winning tariff, while ReNew, one of India’s largest private renewable developers, bid less than 1% above it. The close clustering of bids from two very different and large developers provides further evidence that the winning price was not an outlier. It is consistent with the underlying economics created by rapidly falling solar and battery costs.
How reliable is the project?
The remaining question is whether such a system can maintain the required output during difficult conditions, particularly monsoon periods, unusually cloudy days, and successive days of weak solar generation.
Paliwal et al. (2026) tested this using ten years of hourly weather data across ten Indian states. They find that for every 1,000 MW contracted, they find that a configuration of about 3 GW of solar and 12 GWh of battery storage in Rajasthan can deliver the thermal-mimic profile at an all-in cost below the ₹5.25/kWh auction price. In states with weaker solar resources or stronger monsoon effects, roughly 10–20% more solar is required, while storage remains around 12 GWh; even there, the modeled costs remain within about 7% of the auction price.
Earlier studies had already shown that this type of system could be technically feasible. (Chojkiewicz et al., 2025; Ember; IRENA). What the SECI auction adds is market evidence: major developers are now willing to put binding commercial bids behind that technical proposition. Projects of this scale remain rare globally. For example, Masdar’s 1 GW 24/7 clean-energy project is another prominent example.
A recent CSEP analysis cautions that low storage-auction tariffs can understate the cost of firm power when contracts allow monthly averaging or leave difficult hours to the buyer; under a much stricter every-hour firmness requirement, it estimates costs of roughly ₹8.3–11.8/kWh (Vijay and Tongia, 2026). The thermal-mimic auction addresses much of this concern by specifying output in 15-minute blocks every single day, and with explicit penalties for shortfalls, rather than relying on annual or monthly energy targets.
Three additional advantages not priced in the “thermal-mimic” market test
Recent analysis by Paliwal, Abhyankar and Phadke 2026 explains how this result is achievable and highlights three particularly important advantages beyond just lower prices for comparable performance
1. The cost advantage is significantly understated: ₹5.25/kWh stays fixed for 25 years while conventional power costs rise
The ₹5.25/kWh auction price is not only below the starting price of recently contracted coal power, it is fixed in nominal terms for 25 years. That makes the contract a long-term hedge against fuel-price and freight cost inflation.
Figure 3: Actual and projected utility power purchase costs, average realized revenue of NTPC, recent coal PPA prices, and thermal-mimic auction price
Data sources: PFC, Reports on Performance of Power Utilities, 2009-10 to 2024-25 editions (power purchase cost; FY2010-12 reconstructed from expenditure annexures); NTPC annual reports FY2012-FY2026 (average realized tariff, standalone). Projections: seven TBCB coal PPAs MarNov 2025 (11.9 GW, Rs 5.38-6.30, average 5.81; line anchored on the Rs 5.84 midpoint per SERC adoption orders: UPERC 2228/2025, MPERC 121/2025, WBERC, BERC 36/2025, AERC)
As shown in the figure, India's average power-purchase cost has been rising at 4-5% per year - from about ₹2.7/kWh in FY2010 to ₹5.4/kWh in FY2025 (solid blue line). NTPC's average realised tariff, predominantly reflecting its coal-based generation fleet, increased at a similar rate from about ₹2.6/kWh in FY2011 to ₹4.8/kWh in FY2026.
More importantly, several new coal power purchase agreement (PPA) prices already start above the thermal-mimic price. For example, the seven competitively procured coal PPAs signed in 2025 opened at ₹5.4–6.3/kWh (average of Rs 5.8/kWh). Coal tariffs are not fixed for the contract duration. They contain a fixed cost component (which typically includes depreciation, interest, maintenance etc) and a variable or fuel cost component that escalates over time. Assuming the variable cost increases at 2.6% per year (according CERC tariff norms), the average new coal PPA price could be as high as ₹7.4/kWh by 2050 (dotted red line). The thermal-mimic contract price, by contrast, remains fixed at ₹5.25/kWh throughout (solid golden line). For the 1 GW contract size operating at ~70% capacity factor, this is equivalent to an annual saving of Rs 350 - 1,300 Cr/yr, with a nominal NPV of over Rs 5,700 Cr over 25 years (assuming 10% discount rate).
The difference becomes even clearer in real terms. A nominal tariff fixed at ₹5.25/kWh over 25 years is equivalent to roughly ₹3.79/kWh in real 2026 rupees (assuming 4% annual inflation). In other words, the real cost of power under the contract declines every year. The same is true in dollar terms: ₹5.25/kWh is about \$55/MWh at ₹95 per dollar today; with a 3% annual depreciation of the rupee compared to USD (similar to long-term historical trends), it would fall to roughly \$41/MWh by 2036 and \$27/MWh by the final year of the contract.
The key comparison, therefore, is the fixed price for 25 years versus a coal tariff that starts higher and remains exposed to fuel-cost escalation.
2. Shorter lead times and modularity reduce the costs of overbuilding or underbuilding amid rapid but uncertain demand growth
A second major advantage is the combination of rapid deployment and modularity. The auction requires projects to be commissioned in less than two years, while conventional power plants typically require much longer development and construction periods. The results also show that developers are willing to offer similar tariffs for projects as small as 100 MW, roughly one-tenth the scale of a large coal plant.
Firm solar-plus-storage capacity can therefore be added incrementally as demand materializes. This reduces the risk of committing prematurely to large, indivisible assets that could leave the system with costly excess capacity or supply shortfalls if demand differs from forecasts. Such flexibility is especially valuable given the deep uncertainty surrounding the scale, timing, and location of AI-driven electricity demand.
3. Significant environmental benefits
Carbon emissions impose costs on every country, including the country that emits them. India has contributed relatively little to historical emissions, but it is highly exposed to climate damage. One study estimates India’s domestic social cost of carbon at \$86 per tonne of CO$_2$, with a 66 per cent uncertainty range of \$49 to \$157, the highest central estimate among the countries studied. Assuming coal emissions of 0.9 tonnes/MWh and an exchange rate of ₹95 per dollar, this corresponds to domestic damages of roughly ₹4 to ₹13/kWh, with a central estimate of about ₹7/kWh. These estimates are uncertain, but the policy implication is clear. Even if India disregards the damage its emissions impose on other countries, the avoided damage within India should be included when comparing coal with clean power. (Ricke et al., 2018)
Coal generation also imposes substantial local air-pollution costs. Cropper et al. (2021) estimate that premature mortality caused by air pollution from India’s coal-fired power plants imposes damages of ₹0.73/kWh, using a value of statistical life of ₹10.3 million. The authors describe this as a lower-bound estimate because it includes premature mortality but excludes morbidity and other effects of air pollution, including impacts on neurological development, worker productivity, crop yields, and visibility. Chakravarty and Somanathan (2021) estimate average air-pollution mortality damages from coal generation in India at 2.03 US cents/kWh, equivalent to ₹1.40/kWh using the authors’ 2018–19 exchange rate of ₹69 per USD. A reasonable conservative estimate is therefore that premature-mortality damages alone add roughly ₹1/kWh to the social cost of coal generation in India, with total local air-pollution damages likely higher.
Would it create significant import dependence on China and how to mitigate those risks?
One common criticism of renewable energy plus storage systems is their dependence on Chinese imports, particularly for battery cells. While India has developed a robust solar panel manufacturing base, its battery manufacturing and supply chains remain underdeveloped, and the country is indeed heavily reliant on China. But the relevant question is not simply whether that dependence exists. It is more nuanced such as how large the exposure is, where in the value chain it lies, how much leverage does China have etc.
First, India’s import dependence is concentrated in battery cells. Battery-cell prices have fallen dramatically over the past decade (from roughly \$400–500/kWh in 2015 to around \$50/kWh in 2025) due to the technological progress, manufacturing scale, and substantial excess production capacity in China. As a result, battery cells now account for only about 15-20% of the upfront capital investment in a firm clean-power project (Paliwal et al., 2026).
Their share is even smaller when measured against the project’s full lifecycle cost. Imported cells account for only about 10% of total lifecycle costs. Financing, domestically produced solar equipment, labour, construction, and other storage-system components and services account for the remaining roughly 90%, much of which represents domestic value creation. For example, of the thermal-mimic auction price of Rs 5.25/kWh, the imported-cell component would be only about Rs 0.5/kWh or so. The macroeconomic exposure is thus modest relative to the economic gains.
Second, dependence on imported batteries is different from dependence on imported fuels like coal, oil, or gas. Fuels are consumptive and must be continuously replenished. If supply stops, electricity production can stop. A battery is a capital asset that operates for many years. A disruption in cell imports would affect the construction of new projects, not the operation of existing ones. This significantly reduces the leverage of exporting countries and gives India time to find alternative suppliers, expand domestic production, or modify deployment plans.
The main security risks arise from the electronics and software surrounding the cell. India can import cells while retaining domestic control over pack assembly, inverters, battery-management systems, energy-management systems, firmware, communications, operational data, and remote access. It should also build recycling capacity and maintain some domestic cell manufacturing, as it has sought to do with solar equipment.
What are the implications? What more needs to be done?
First, governments and utilities should not commit to large fleets of new coal and gas plants without allowing renewables and storage to compete against the same performance requirements. This does not imply that renewables and storage will win in every location or for every operating profile. It means that the presumption in favour of coal and gas is no longer justified. All-source competition should determine which portfolio can provide the required reliability at the lowest cost.
Second, this competition must compare full costs. A fixed-price contract has value when fossil-fuel costs are exposed to inflation. Modularity has value because utilities can procure capacity in smaller increments as demand emerges, reducing the risks of overbuilding and underbuilding. Short lead times also have value when demand is growing but uncertain. Environmental damage should be priced rather than treated as free. Utilities should specify the quantity, operating profile, and commissioning date they require. Any resource that meets these requirements at the lowest total cost should win.
Third, these changes weaken the case for monopoly utility procurement. Large, slow, and scale-intensive power plants once favoured centralised planning and a single buyer backed by a distribution monopoly. Firm, round-the-clock power can now be assembled from modular solar and storage projects with much shorter lead times. The wires network remains a natural monopoly because duplicating distribution infrastructure is wasteful. Power generation, procurement, and retail supply do not require the same monopoly. Competitive suppliers can buy and sell power, while distribution utilities operate the network, provide last-resort service, and protect small and vulnerable consumers. ERCOT shows that competitive markets can support rapid investment in renewables and storage, although its design cannot simply be copied. Prayas (Energy Group) has outlined pathways for India, while proposed amendments to the Electricity Act also point towards greater competition. In the United States, PJM and CAISO should examine which elements of the ERCOT model, including faster interconnection and stronger market signals, can be adapted to their systems. GridLab’s recent work provides one such pathway. A wires-focused utility may also become financially stronger as electrification expands demand for network services.
Fourth, the gains from low-cost clean power extend far beyond the electricity sector. Time-varying prices can encourage flexible consumers to use electricity when clean supply is abundant and inexpensive. This can make industrial heat, hydrogen, steel, aluminium, transport, and other activities cheaper to electrify or decarbonise. Power-market reform is therefore not merely an electricity-sector reform. It can provide the foundation for lower-cost industrialisation and a cleaner economy.
References
Assam Electricity Regulatory Commission (AERC) (2025). Order dated 22 October 2025 on procurement of 3,200 MW of coal-based thermal power by APDCL. Tariff approval for the Assam thermal-power procurement cited in Figure 3. Cited in MPERC, Order in Petition No. 121/2025, paragraph 73.
Bihar Electricity Regulatory Commission (BERC) (2025). Order in Case No. 36/2025: Adoption of tariff for long-term procurement from the 2,400 MW Pirpainti Thermal Power Station. Final order, 27 August. Patna: BERC.
Chakravarty, S., and E. Somanathan (2021). There is no economic case for new coal plants in India. World Development Perspectives, 24, 100373. DOI: 10.1016/j.wdp.2021.100373.
Chojkiewicz, E., N. Abhyankar, U. Paliwal, and A. Phadke (2025). Declining costs make solar plus storage economical for industrial captive power. iScience, 28(11), 113763. DOI: 10.1016/j.isci.2025.113763.
Cropper, M., R. Cui, S. Guttikunda, N. Hultman, P. Jawahar, Y. Park, X. Yao, and X.-P. Song (2021). The mortality impacts of current and planned coal-fired power plants in India. Proceedings of the National Academy of Sciences, 118(5), e2017936118. DOI: 10.1073/pnas.2017936118.
International Energy Agency (IEA) (2026). Technology: Solar PV and wind. In Global Energy Review 2026. Paris: IEA.
International Renewable Energy Agency (IRENA) (2026). 24/7 renewables: The economics of firm solar and wind. Abu Dhabi: IRENA, May.
Madhya Pradesh Electricity Regulatory Commission (MPERC) (2025). Order in Petition No. 121/2025: Adoption of tariff for long-term procurement of 3,200 MW plus 800 MW under the greenshoe option from new power stations in Madhya Pradesh. Final order, 15 December. Bhopal: MPERC.
NTPC Limited (various years). Annual reports. FY2012-FY2026 editions cited in Figure 3 for standalone average realised tariffs. Report collection.
Paliwal, U., N. Abhyankar, and A. Phadke (2026). Closing the Credibility Gap: Solar-Plus-Storage Delivers Coal-Equivalent Reliability at Lower Cost in India. Working paper, March. India Energy and Climate Center, Goldman School of Public Policy, University of California, Berkeley.
Paliwal, U., A. Phadke, and N. Abhyankar (2026). India's Renewable Energy Breakthrough: Coal-Like Reliability at a Lower Fixed Price. Working paper, August. India Energy and Climate Center, University of California, Berkeley.
Power Finance Corporation (PFC) (various years). Report on Performance of Power Utilities. 2009-10 to 2024-25.
Ricke, K., L. Drouet, K. Caldeira, and M. Tavoni (2018). Country-level social cost of carbon. Nature Climate Change, 8, 895-900. DOI: 10.1038/s41558-018-0282-y.
Solar Energy Corporation of India (SECI) (2026). Request for selection: 1,000 MW firm and dispatchable renewable energy round-the-clock power (SECI-FDRE-RTC-V). Tender SECI000240; SECI/C&P/IPP/13/0020/25-26, 10 March, with amendments.
Uttar Pradesh Electricity Regulatory Commission (UPERC) (2026). Order in Petition No. 2228/2025: Approval of the power supply agreement and adoption of tariff for procurement of 1,500 MW from Mirzapur Thermal Energy (UP) Private Limited. Final order, January. Lucknow: UPERC.
Vijay, R., and R. Tongia (2026). Electricity Storage is Getting Quite Cheap - But Firmer Storage Isn't as Cheap as Sometimes Believed. Centre for Social and Economic Progress (CSEP), 7 September.
West Bengal Electricity Regulatory Commission (WBERC) (2025). Order in Case No. OA-513/24-25: Adoption of competitively discovered tariff for 1,492 MW contracted capacity from a new 2 x 800 MW greenfield thermal power plant. Order, 28 May. Kolkata: WBERC.
The authors are researchers at India Energy and Climate Center, University of California, Berkeley
