Ask a cement plant, a steel mill and a foundry the same question — what does electricity cost you? — and you will get three answers that sound similar and mean completely different things.
All three are what the Bureau of Energy Efficiency calls designated consumers: industries whose annual energy use is large enough that the government tracks and now regulates it directly. All three buy power in quantities where a paisa per unit is a real number on the profit and loss account. And all three are being pushed, at the same time, by tariffs that have become time-sensitive and by obligations that have become legally binding.
But their load shapes are nothing alike. And load shape, more than sector, is what decides whether a battery earns anything on an industrial site.
The scale of the exposure
Start with how much electricity each one actually buys.
For cement, widely cited industry figures put electrical consumption at roughly 120 kWh per tonne of cement, rising to around 150 kWh per tonne of clinker. Cement’s total energy bill is dominated by thermal fuel for the kiln, but the electrical share — grinding, the raw mill, the kiln drives, the packing plant — is large, continuous and entirely grid-facing.
For steel via an electric arc furnace (EAF), published ranges sit at roughly 350 to 600 kWh per tonne of liquid steel, depending on scrap quality, charge mix and how hard the furnace is pushed. Here electricity is not a support service. It is the process — the arc is doing the melting.
For foundries, BEE cluster studies are the most useful source because they measured real Indian units rather than modelled them. An induction-furnace foundry consumes about 1,000 to 1,200 kWh per tonne of good castings, of which roughly 600 to 700 kWh per tonne is the melting itself and the remainder is everything else — moulding, sand plant, compressed air, handling. In such a unit, electricity accounts for 85 to 95 per cent of total energy consumption, the induction furnace alone draws 70 to 85 per cent of the electrical energy, and energy works out at roughly 25 to 30 per cent of manufacturing cost. Around 70 per cent of Indian foundries have now moved to induction melting.
Hold that last figure in mind. When a quarter to a third of your cost of manufacture is electricity, an arrangement that shaves a few per cent off the power bill is not a sustainability gesture. It is margin.
Cement: a flat load with a renewable problem
A cement plant runs continuously and its electrical load barely moves. That is unhelpful news for anyone selling peak shaving, because there is no peak to shave — the plant’s maximum demand is close to its average demand, and a battery that flattens nothing saves nothing on the fixed charge.
The cement sector also got there first on self-generation. Waste heat recovery (WHR) — capturing heat from the preheater and clinker cooler exhaust to run a steam turbine — is now standard practice rather than an innovation. The cement industry has implemented around 538 MW of WHR capacity, the largest of any Indian sector, concentrated in Rajasthan (roughly 42 per cent), Chhattisgarh (about 16 per cent) and Madhya Pradesh (about 11 per cent), which together hold some 70 per cent of installed capacity. MNRE’s 2016 assessment put the sector’s technical potential at around 1,100 MW, so there is still headroom.
So where does a battery fit? Almost entirely on the renewable side, not the load side.
Cement’s binding problem is that it must now consume a rising share of non-fossil energy while running three shifts. Solar generates for a fraction of the day. Wind is better spread but unreliable hour to hour. WHR is steady but tied to kiln operation — when the kiln comes down for maintenance, the WHR output goes with it. Storage is what converts a contracted renewable profile into something that actually lands inside the plant’s own consumption instead of being exported at a poor rate. That is the renewable firming duty, and it is why the newest cement power deals increasingly bundle a battery with the solar.
Steel: the violence of an arc furnace
An electric arc furnace is one of the most disruptive loads connected to any distribution network. Melting scrap with an arc produces large, erratic fluctuations in reactive current, and the literature on this is consistent: arc furnaces introduce harmonics, inter-harmonics, voltage flicker at frequencies up to around 25 Hz, voltage dips and phase unbalance into the supply. India ranks among the world’s largest producers of steel through electric melting, so this is not a niche concern.
Two consequences follow.
The first is commercial. Flicker and unbalance are things utilities measure and, in many jurisdictions, penalise or require you to mitigate at your own cost before granting or retaining a connection at the capacity you want. A mill that cannot hold its power quality within limits may find its expansion constrained by the grid rather than by its own equipment.
The second is operational. A voltage dip mid-heat is expensive in a way that is easy to underestimate — not just the lost energy, but a furnace-full of metal at the wrong temperature and a delayed cast.
Storage does not solve arc furnace power quality on its own; the conventional answers are static VAr compensators and D-STATCOMs, and those remain the workhorses. But a battery paired with fast power electronics can supply and absorb real power on a sub-second timescale, which is precisely the timescale of the disturbance, and hybrid arrangements are an active area of deployment. On the plainer commercial side, a mill also has the same demand-charge and tariff exposure as any other large consumer — the mechanics of which we set out in demand charges explained.
Foundries: where the arithmetic is cleanest
Of the three, the foundry is where a battery most often pays for itself on straightforward, defensible numbers.
An induction furnace melts in batches. The furnace draws hard while a heat is coming up, then drops back. In a shop with several furnaces, the billed maximum demand for the entire month is frequently set by the few intervals when two or more happened to be drawing at once — a scheduling accident, not a production requirement. Most Indian industrial tariffs bill a fixed charge on that highest recorded demand, so the plant then pays all month for something that happened for fifteen minutes.
This is the textbook case for peak shaving. A battery sized to the overlap, not to the whole load, discharges during those coincident intervals and caps the recorded maximum. Because the energy involved is small and the saving applies to the whole billing period, the arithmetic is unusually favourable — and unusually easy to verify from your own meter data before spending anything.
Two other foundry-specific points are worth noting. Disciplined melt scheduling is the free version of the same fix, and should be tried first — if you can stagger heats so that furnaces never peak together, you have achieved part of the result at no capital cost. And an interruption mid-heat in an induction furnace risks a solidified charge, which is a genuinely costly event, so the ride-through value here is real in a way it is not for, say, a grinding operation that can simply restart.
The three side by side
| Cement | Steel (arc furnace) | Foundry (induction) | |
|---|---|---|---|
| Load shape | Near-flat, continuous, three shifts | Violently fluctuating, second by second | Batch melt cycles with idle periods |
| Electrical intensity | ~120 kWh per tonne cement | ~350–600 kWh per tonne liquid steel | ~1,000–1,200 kWh per tonne castings |
| Peak vs average demand | Close together — little to shave | High and erratic | High during coincident melts |
| Existing self-generation | Waste heat recovery is standard | Limited; some captive thermal | Rare |
| Primary battery value | Firming captive renewables | Power quality and ride-through | Demand charge reduction |
| Secondary value | Peak-window tariff arbitrage | Demand charges, process continuity | Interruption protection mid-heat |
| Typical scale | Containerised, multi-MWh | Containerised, sized to disturbance | Cabinet to single-MWh |
| Hardest part of the case | Flat load means no demand saving | Mitigation alternatives already exist | Persuading a small unit to meter first |
The policy squeeze is now the bigger driver
For most of the last decade, storage in heavy industry was a cost-saving conversation. It has become a compliance conversation, and that is a more forceful one.
Renewable Consumption Obligation (RCO). Under the Energy Conservation (Amendment) Act, 2022, designated consumers — explicitly including cement and iron and steel — must meet a minimum share of their electricity consumption from non-fossil sources. The trajectory is 33.01 per cent in FY2025-26, 35.95 per cent in FY2026-27, and rising to 43.33 per cent by FY2029-30. BEE is the nodal authority for monitoring and compliance. This is the single most important number for a plant manager reading this, because a three-shift operation cannot meet a 35 per cent non-fossil share from daytime solar alone without either storage or a very large contracted wind position.
Carbon Credit Trading Scheme (CCTS). Final emission intensity targets for the first four sectors were notified on 8 October 2025, covering aluminium, cement, chlor-alkali and pulp and paper. For cement the required reduction is roughly 4.7 to 7.6 per cent, measured against an FY2023-24 baseline, across compliance years FY2025-26 and FY2026-27, with the reduction back-loaded — roughly 40 per cent of it in the first year and 60 per cent in the second. Targets for further sectors followed, and a draft notification introducing plant-wise targets for iron and steel was published in mid-2026. Across all nine sectors, something in the order of 740 entities are expected to carry legally binding intensity targets.
Green Steel Taxonomy. The Ministry of Steel notified India’s green steel definition by Gazette notification in December 2024. Steel qualifies as green below 2.2 tonnes of CO₂-equivalent per tonne of finished steel, with a star rating above that threshold: five star below 1.6, four star between 1.6 and 2.0, and three star between 2.0 and 2.2. The National Institute of Secondary Steel Technology handles measurement, reporting and verification and issues the certificates, and the thresholds are to be reviewed every three years. Because an arc furnace’s emissions are largely the emissions of the electricity feeding it, the grid mix a mill buys from now directly determines what rating its product can claim.
Put those three together and the picture is clear enough. An energy-intensive plant now has a legal requirement to consume more non-fossil power, a legal requirement to lower emission intensity, and — in steel — a commercial reason to prove how clean its input electricity was. All three point at contracting renewable generation. And contracting renewable generation for a continuous industrial load is where storage stops being optional.
Why 26 per cent keeps appearing
The structure that makes this work in India is group captive, and it explains a pattern that looks strange from outside: cement companies buying exactly 26 per cent of renewable project companies, over and over.
Indian electricity law treats a consumer as captive — and therefore exempt from cross-subsidy surcharge and additional surcharge — if it holds at least 26 per cent of the equity in the generating project and consumes at least 51 per cent of the electricity generated. Those surcharges are a substantial part of the delivered cost of open access power, so clearing the threshold materially changes the tariff. Hence 26 per cent, almost never 27.
The recent deal flow shows what this looks like in practice. UltraTech Cement has taken 26 per cent positions in several project companies including a 91 MWp DC solar plant in Chhattisgarh, and reported green energy reaching 35.8 per cent of its power mix after commissioning 371 MW of renewable capacity in FY2026. Dalmia Bharat took 26 per cent of a hybrid project combining 21.6 MW of wind with 14 MWp of solar for its Kadapa plant. And JK Lakshmi Cement approved an investment for 26 per cent of a project pairing a 29 MW AC solar plant with a 28 MWh battery energy storage system to supply its Sirohi plant in Rajasthan.
That last one is the tell. The battery is there because a cement plant runs at night and a solar plant does not. We cover how the captive and open access routes compare, and where each one breaks, in captive versus open access for factories and in the green open access rules and storage.
Alongside all of this sits the plainer tariff story. The Electricity (Rights of Consumers) Amendment Rules, 2023 made time-of-day pricing mandatory, with peak-period tariffs for commercial and industrial consumers set at at least 1.20 times the normal rate and solar-hour rates set below it. For a plant that can move any load at all, that spread is money. For one that cannot, a battery moves it instead.
Policies, obligation trajectories and tender terms in this area change by notification, and several of the instruments above — the CCTS steel targets in particular — were still moving through draft stages during 2026. Verify the current position with BEE, the Ministry of Power and your State Commission before committing capital on the basis of any figure here.
What this means for you
If you run a foundry or a small melting shop, this is the most actionable case on the list and the cheapest to test. Pull twelve months of half-hourly data from your meter and find the intervals that set your billed maximum demand each month. If a handful of coincident melts are doing it, you have a well-defined problem with a well-defined answer. Try the free fix first — stagger the heats — and size a battery only to the overlap that remains. A cabinet-scale system such as the ADESS 250 is usually the right order of magnitude here; you are buying power for minutes, not hours.
If you run a cement plant, stop looking for a behind-the-meter saving and start with your RCO position. Work out what share of your consumption is non-fossil today, what the FY2026-27 requirement of 35.95 per cent implies in absolute terms, and how much of a contracted solar or wind profile would actually land inside your load without storage. The gap between those two numbers is your storage requirement. At that scale you are looking at containerised systems in the ADESS 5000 class rather than cabinets, and the commercial structure — group captive at 26 per cent, or a third-party developer arrangement — matters as much as the hardware.
If you run a steel mill, split the question in two before anyone quotes you anything. The power quality problem and the energy cost problem have different answers, different suppliers and different economics, and conflating them is how these projects go wrong. Get a power quality study at the point of supply first. If flicker and unbalance are constraining your connection or costing you penalties, that is an engineering problem where storage is one option among several. Separately, if you sell into markets that care about the Green Steel Taxonomy rating, the carbon intensity of your purchased electricity is now a product specification rather than an overhead — and that argues for contracted renewables with firming regardless of what the power quality study says.
If you are a lender, promoter or CFO looking across a portfolio of heavy industrial assets, the useful reframing is this: energy is no longer purely an operating cost line. Under CCTS it is a compliance liability with a price attached, and under the Green Steel Taxonomy it is a market access question. A plant with a credible path to its FY2029-30 obligation is a materially different asset from one without.
Wherever you sit, the sequence is the same and it starts with data rather than equipment: meter first, find your real peak-setting intervals and your real non-fossil share, then size. Our industry-by-industry guide to storage use cases is a good place to see how other sectors approached the same question. To put your own tariff and load profile through the numbers, use our BESS savings calculator — or if you would rather walk a site engineer through your bill and your meter data, talk to our team.
Policy snapshot as of September 2026. RCO trajectories, CCTS sectoral targets, Green Steel Taxonomy thresholds and state time-of-day tariff windows are revised by notification; confirm current terms with BEE, the Ministry of Power, the Ministry of Steel and your distribution licensee before investment decisions.