India is adding clean power faster than its electricity system can make it affordably and reliably available to consumers.
Between April and June 2026 alone, India curtailed or restricted 8,133 GWh—more than eight billion units—of solar generation. Clean electricity was available, but the grid could not absorb and deliver it.
This is India’s power-transition challenge captured, perhaps, in one number. Contrary to conventional wisdom, rapid capacity additions and low auction tariffs do not necessarily ensure that consumers receive affordable, clean electricity when and where it is needed.
The issue is not only cheap solar or wind generation. It is whether generation, transmission, storage, flexible capacity, grid operations and markets advance as one coordinated system. Otherwise, India can add low-cost clean capacity while raising the cost at the meter through curtailment, congestion, balancing and backup.
India must deliver electricity that is simultaneously affordable, abundant, reliable and clean—24×7. This is very different from equating cheap, clean generation with cheap, clean electricity. It requires fundamental change in grid planning and operations and in how markets value the resources and services that electricity delivery requires.
Traditional planning forecasts demand, determines the required supply assets, builds around that forecast and revisits the plan periodically. Operators react when demand, weather or plant availability differs from expectations. This worked when demand was predictable and most generation could be controlled. Neither assumption holds any longer.
Solar and wind output changes with the weather—sometimes dramatically during prolonged wind lulls and cloudy monsoon periods. Demand is also harder to extrapolate from history. Electric mobility, industrial electrification and data centres are changing when, where and how electricity is consumed. Major computing and industrial clusters can add concentrated loads faster than conventional planning cycles can accommodate.
One essential pillar for matching demand and supply under this uncertainty is an adaptive electricity system—one planned continuously and operated intelligently. Traditional planning selects a future and builds around it. Adaptive planning prepares for several possible futures and revises decisions as reality unfolds.
The intelligence enabling this adaptation is physical AI for the grid—not the LLM-based generative AI dominating public attention. It combines weather, demand, plant and network data, market information and engineering models to assess possible conditions and compare physically feasible actions. Planning and operations become a continuous loop: observe, anticipate, act, learn and adapt.
At noon, solar output may surge and then fall as evening demand rises. The system must decide whether to charge storage, shift consumption, adjust conventional generation or preserve reserves, while delivering electricity at the lowest practicable cost and emissions without compromising reliability. Physical AI compares the options, operators decide, and the system learns from the outcome.
Operating evidence must also guide grid investment. When a bottleneck recurs, physical AI can test whether the next rupee creates greater value in transmission, storage, flexible generation or demand response. Its output is not prose; it is a better operating or investment decision. It must remain grounded in grid physics, explainable to operators and under human control.
Adaptive planning and intelligent operations form one pillar. Market design forms the other. An adaptive grid can identify what the system needs; markets must translate those needs into investment and operating behaviour. A market that primarily rewards electricity generated will produce energy, but not necessarily dependable generation during stressed periods, rapid ramping, adequate storage duration, reserves, congestion relief or flexible demand. These services have measurable value and must be procured and compensated. Prices and contracts— including options for capacity and flexibility—should reveal when and where flexibility is scarce, allowing generators, storage providers and consumers to respond.
Neither pillar is sufficient without the other.
Adaptive planning without appropriate incentives cannot mobilise the required resources. Market reform without adaptive planning risks pricing yesterday’s requirements instead of anticipating tomorrow’s conditions.
A low solar or wind tariff is therefore not the cost of electricity at the meter. When generation, networks and flexibility are developed separately, consumers also pay for curtailment, underutilised firm capacity, balancing and delayed network reinforcement. As variable renewable penetration increases, these costs can compound. An adaptive, intelligently operated grid prevents low-cost variable generation from being offset by escalating reliability and grid-stability costs.
The Draft National Electricity Policy 2026 recognises many necessary building blocks: resource adequacy, flexible transmission, storage, capacity markets, ancillary services, demand response, interoperable data and AI. Its opportunity is to connect them within one coherent framework.
Generation, transmission and flexibility must be planned together. Future adequacy must be continually reassessed using operating evidence. Grid institutions need interoperable data and a shared operating picture. Markets must reward dependable performance and system value rather than technology labels. Expandable, interconnected grid corridors can support major data-centre and industrial clusters, with infrastructure added as demand materialises.
India’s power transformation will not be judged by gigawatts of solar or wind, installed or auction tariffs announced. It will be judged at the meter. Affordable, abundant, reliable and clean electricity is the objective. Adaptive planning and intelligent operations, together with markets designed around actual system value, are how India can deliver it.