The Net-Zero Arithmetic: The 2070 Question
yet naming is the least part of the climb.
The first miles run downhill, and flatter false;
the sheer of it is saved for the last time.”
India’s net-zero question is usually framed as a capacity race, and by that measure India is winning — it crossed 50% non-fossil capacity five years early. But capacity is not the constraint. The constraint is generation, and the shape of the descent after it. India still burns coal for roughly 70% of its actual electrons; its demand is set to more than double this decade; and the hard part of any 2070 pathway — the near-total collapse of coal generation — is entirely back-loaded past 2040. The arithmetic, not the ambition, is where the answer lives.
The gap that defines everything
Start with the single fact most coverage blurs: capacity share is not generation share. Coal is about 43% of India’s installed capacity but roughly 71% of the electricity actually generated. Non-fossil sources are about 53% of capacity but only around 29% of generation. The reason is load factor: coal runs at ~65% of nameplate around the clock, while solar delivers at ~20% and only in daylight. Five hundred gigawatts of solar nameplate does not displace five hundred gigawatts of coal generation — it displaces the fraction that shows up when the sun is out.
Share of installed capacity vs share of electricity generated, India (~2026 capacity; FY25-26 generation). The residual is gas and other. Source: CEA / MNRE.
Coal is ~71% of India’s generation but only ~43% of its capacity. The distance between those two numbers is the entire net-zero problem — and the 50%-non-fossil-capacity milestone measures the easier one.
This is why the capacity milestone, real and early as it is, understates the task. It measures nameplate, which India adds faster than almost anyone. Net zero is measured in electrons and emissions, and there the coal share has barely moved — from ~71% to ~70% on the last peak day. The build is impressive; the displacement is slow.
What net zero would actually take
To retire coal generation — not merely dilute coal’s share — you need either firm clean power or a very large overbuild of intermittent power plus storage. This is where the arithmetic gets steep. Solar is effectively unlimited and will do the bulk of the work, but solar alone cannot replace coal’s role, only its daytime energy. The genuinely under-exploited levers are the firm-clean ones: nuclear, at just 8.8 GW today, targeted at ~22 GW by 2035 and 100 GW by 2047; and Himalayan hydro, with ~76 GW identified but locked behind transmission and land. These matter disproportionately because only firm clean power substitutes for what coal actually does.
The overbuild math explains the headline numbers. Offsetting ~1,300 billion units of coal generation with ~20%-load-factor solar needs roughly five times the nameplate, plus storage to move daytime energy into the evening. That is why the canonical pathway lands where it does: solar rising to ~5,630 GW by 2070 from ~162 GW today (a ~35-fold increase), coal generation peaking around 2040 and falling 99% by 2060, and 174 GW / 888 GWh of storage envisaged by 2035-36 alone. The endpoint fleet has nuclear heading toward ~300 GW, coal down to ~4% of the mix from ~59% in 2023, fossil fuels at 5-6% of primary energy, and electric vehicles at 84% of the car fleet — at a cumulative cost near $21 trillion, roughly 4.1% of GDP in the net-zero year.
The demand treadmill
What makes the descent hard is that the denominator will not sit still. India’s electricity demand more than doubles this decade: from ~1,840 billion units and a ~250 GW peak now to ~3,365 billion units and a 459 GW peak by 2035-36, an energy CAGR of 6.4% on the planners’ assumptions — and recent actual growth has run nearer 9%. The drivers are precisely the markers of a consumerist take-off: 50-plus GW of data-centre load by 2030, electric vehicles adding 80-100 TWh, green-hydrogen electrolysers, an air-conditioning boom, and urbanisation adding an Australia’s worth of city-dwellers every 18 to 24 months. Electricity’s share of final energy rises from ~21% today toward 40-60% by 2070.
Electricity demand vs coal generation on a net-zero-consistent path. Demand soars while coal must peak ~2040 and fall ~99% by 2060. Illustrative: CEA to 2035-36, CEEW pathway shape thereafter.
The chart is the whole problem in one frame. The gap that opens between the two lines — demand climbing, coal peaking and then collapsing — is what firm clean power and storage must fill, on schedule, for decades. And it is filling slowly. A corollary worth stating plainly: electrifying end-uses such as vehicles and cooking helps the fuel-import bill and raises efficiency, but on a ~70%-coal grid it mostly moves emissions upstream. It is a bet on generation decarbonising, not a substitute for it.
The both/and, and the lock-in
India is running an explicit both/and. Alongside ~158 GW of renewables under construction, it has ~97 GW of new coal planned through 2034-35; coal capacity rises to ~315 GW by 2035-36 even as its share falls to 28%. The near-term logic is sound: coal is the firm, dispatchable anchor for non-solar hours and seasonal renewable dips, and storage is not yet available at the scale required to replace it.
The risk is lock-in. A coal plant commissioned in 2030 has a 30-to-40-year life — it is designed to run into the 2060s, exactly when a 2070 pathway requires coal generation near zero. The entire strategy rests on a wager that this new coal runs as low-load-factor backup rather than baseload. If it runs as backup, the carbon budget survives and the plants are insurance. If it runs as baseload — because demand outran clean supply, as it has every year so far — then it either blows the budget or becomes a stranded asset before its debt is repaid. That wager is unproven at this scale, and it is the hinge on which the back half of the pathway turns.
Is 2070 reasonable, or will it slip?
First, a correction to a common premise: 2070 has not been postponed. India named it at COP26 in 2021 — having pointedly refused to set any date before — and it was already the latest target among major economies, two decades behind the US and EU (2050) and China (2060). In March 2026 India raised its interim ambition, to a 47% emissions-intensity cut and 60% non-fossil capacity by 2035. The direction of travel is more ambition, not less. So the question is not whether the label slips.
The question is whether the peak and the descent hold, and there the picture splits cleanly. The front end, to about 2035, is credible and probably beaten: India over-delivers on capacity targets and the economics of new solar and wind are firmly with it. The back end, 2040 to 2070, is where the genuine risk sits. Absolute emissions are still rising — India’s target is intensity-based, not an absolute cap, and emissions were ~4.4 Gt in 2024 and climbing; the coal build wave locks generation in past 2050; demand keeps outrunning the planning assumption; and nuclear and storage have a long history of missing their targets.
India is more likely to hold 2070 formally than to hit its implied trajectory. The real risk is not that India postpones the date — it will not need to; a target that distant absorbs slippage — but that the emissions peak drifts from ~2040 toward ~2045-2050, forcing a steeper, costlier, finance-dependent descent that may not materialise. In one line: expect India to hit its capacity targets early and its emissions targets late — the capacity-versus-generation gap, written across five decades.
And the pathway is conditional in a way the headline date is not. It depends explicitly on concessional climate finance from developed economies — roughly $21 trillion in total, about 4.1% of GDP in the net-zero year and nearer 7% if the target were preponed to 2050 — finance that has not materialised at scale. Absent it, the descent slows regardless of what the label says. 2070 is reasonable precisely because it was chosen to be: distant enough to be safe, and contingent enough to be deniable.
The investment read
The RE build is triple-underwritten
Renewables are backed by economics (cheapest new power), by the balance of payments (import substitution), and by capacity targets India keeps beating. This is the sturdiest leg of the whole story and the one least dependent on the 2070 date holding.
Coal is not dead near-term
~315 GW by 2035-36 and a firm-power premium keep coal earning through the 2030s. But the 30-40 year asset life against a post-2040 collapse is real stranded-asset and lock-in risk — own the cash flows, price the tail.
Firm-clean and storage are the alpha
Nuclear, pumped hydro, batteries, and transmission are where the pathway is capital-constrained and where the binding constraint actually sits. The scarce, hard-to-replicate assets are firm clean power, not more nameplate solar.
Energy Security as a Luxury Good (G17) — the same treadmill: domestic supply cannot yet outrun demand, so independence waits.
The Import Bill (G16) — the balance-of-payments case that underwrites the RE build independent of the climate target.
The Cost-of-Capital Gap (G10) — the $21-trillion price tag runs straight into the WACC penalty; finance is the binding constraint on the descent.
Committed Emissions — the lock-in logic in its general form: what today’s fleet obliges tomorrow to emit.
India will almost certainly reach net zero later than it decarbonises its capacity mix, because the two were never the same thing. The 2070 date is safe precisely because it is distant; what is uncertain is the year emissions peak and the slope of the fall after it. The capacity race is being won in public; the generation race is being run quietly, and much more slowly.
For an investor, the durable reads are the renewables build (underwritten three ways over), the scarcity of firm clean power (nuclear, hydro, storage), and a healthy skepticism toward new coal as anything but a bridge. The number to watch is not installed capacity — India will keep beating that — but coal generation, and the year it finally peaks. Until that line turns down, net zero is a direction, not yet a descent.
it yields to feet, and only feet, and time.
A date is but a promise made to distance —
kept, in the end, by those who choose to climb.”
Leave a Reply