Why cement exists, and what the customer buys
Cement is 10–16% of concrete by mass, 20–37% of its cost, and a rounding error on the price of a building. That combination — a product the buyer cannot verify, cannot easily substitute, and barely notices the price of — is why cement is a regional oligopoly everywhere on earth rather than a global commodity. Everything the plant does is downstream of it.
On this page · 17 sections
- 01The one idea
- 02The parts — what cement replaced
- 03First principles — carbonation and hydration
- 04What must be true
- 05The quantitative anatomy — what the customer is paying for
- 06What goes wrong here
- 07Scenarios from the field
- 08The numbers that matter
- 09Build the model
- 10Upgrade paths — changing what you sell
- 11The frontier, and who is running it
- 12What's changing now
- 13Check yourself
- 14Mini case
- 15Go deeper
- 16Carry forward
- 17Where this connects
§ 01 The one idea
Every civilisation has needed to turn loose stone into solid structure. Cement is the answer that works everywhere, and its functional job fits in one sentence:
Convert cheap local rock plus heat into a powder that turns loose aggregate into artificial stone — in bulk, under water, on a construction schedule, using a chemistry that is indifferent to where the aggregate came from.
The last clause is the commercially important one and it is the one outsiders miss. The aggregate is always local, always cheap, and always 80–85% of the mass. Cement is the small, transportable, manufactured fraction that makes the large, local, worthless fraction structural. It is a universal binder in the literal sense: the same powder works on granite chippings in Karnataka and river sand in Bihar.
That universality is why the industry is enormous. Roughly 30 billion tonnes of concrete are used each year, making it the second most consumed material on earth after water, and cement production runs at about 3.8 billion tonnes a year.
"Concrete is the most-used man-made material on earth" is defensible. "The most-used material on earth" is not — water outranks it. Nature's 2025 framing is "second most consumed material globally after water." A consultant who overstates this in a first meeting has spent credibility on a fact they did not need.
§ 02 The parts — what cement replaced
Cement did not arrive into a vacuum. It displaced a family of binders, each of which fails in a specific and instructive way. Understanding how they fail is the fastest route to understanding what Portland cement actually does.
| Binder | Sets by | Under water? | In bulk? | Time to useful strength | Why it fails as a structural binder |
|---|---|---|---|---|---|
| Clay | drying | no | yes | hours | Re-softens when wet. No strength. |
| Gypsum plaster | hydration | no | yes | ~30 min | Soluble in water; creeps under sustained load. |
| Air lime | carbonation — needs atmospheric CO₂ | no | no | months to years | CO₂ cannot reach the interior of a thick section. Will not set under water at all. |
| Hydraulic lime / natural cement | hydration | yes | yes | days–weeks | Limited strength, and dependent on finding a specific marl deposit. |
| Portland cement | hydration of calcium silicates | yes | yes | 1–28 days | — |
§ 03 First principles — carbonation and hydration
Two reactions separate lime from cement, and the difference between them explains the entire modern construction industry. This is worth ten minutes even if you never studied chemistry.
The lime cycle — a closed loop that cannot set in the middle
Burn limestone and you drive off carbon dioxide: CaCO₃ → CaO + CO₂, at around 900 °C. Slake the quicklime with water and you get Ca(OH)₂. Spread it, and over months to years it pulls CO₂ back out of the air and turns into limestone again: Ca(OH)₂ + CO₂ → CaCO₃ + H₂O.
Notice what that means. The carbon you drove off comes back. Air lime is, in principle, a closed carbon loop. It is also useless for anything thick, because the CO₂ has to physically diffuse in from the surface — and diffusion through an already-hardened rind gets slower the further it has to go. The middle of a thick lime wall can stay soft for decades. Under water there is no atmospheric CO₂ at all, so it never sets.
Hydration — setting from within
Portland cement's principal reaction needs no atmosphere: C₃S + H₂O → C-S-H + Ca(OH)₂. The water is already distributed through every part of the mix when it is placed. The binder forms simultaneously throughout the section, in bulk, and under water — because water is the reactant, not an obstacle.
The load-bearing product is C-S-H, calcium silicate hydrate, an ill-crystallised gel that has been the load-bearing phase of every concrete structure since the 1840s.
Air lime gives its carbon back. Portland cement does not. The CO₂ driven off during calcination is released permanently — only a slow, partial reabsorption happens as the concrete carbonates over decades. That asymmetry is the origin of cement's process-emissions problem, and it is chemistry rather than engineering: roughly 60% of cement's CO₂ comes from the calcination reaction itself, not from the fuel. No kiln improvement touches it.
Now look at the right-hand equation again. Hydration produces C-S-H and a large amount of calcium hydroxide, which contributes almost nothing to strength. A pozzolan — fly ash, slag, calcined clay, or Roman volcanic ash — consumes that spare Ca(OH)₂ and converts it into more C-S-H. That is the entire mechanism of blended cement, and it is the same reaction the Romans used. Mechanistically, PPC is the descendant of Roman concrete; OPC is not.
§ 04 What must be true
Cement is bought by people who cannot test it, used by people who did not buy it, and judged weeks after it was poured. That gap between purchase and verification is what the whole product specification exists to close. These conditions are what a cement has to satisfy for the business to work, and none of them has changed in a century.
| What must be true | Why | Target | How you verify it |
|---|---|---|---|
| 1. It gains strength on a schedule the site can work to | A contractor strips formwork on a calendar, not on a test result. Early strength is what is actually bought; 28-day strength is what is specified. Blended cements shift the two apart, and the complaint is always about the first. | grade strength at 3, 7, 28 d |
Ask for the strength distribution, not the mean. A plant can meet a grade on average and fail a customer on a Tuesday. |
| 2. It behaves the same bag to bag | Consistency is worth more to a customer than any single property, because a mix design is calibrated once and then repeated. A cement that moves forces the customer to re-trim water and admixture, which they cannot do at scale. | low SD, not a high mean |
Strength standard deviation month on month, and the plant’s own rejection record. This is the quality metric customers actually act on and the one least often reported. |
| 3. It sets neither too early nor too late | Setting time is the one property that can lose a customer a day’s work. It is controlled by the sulphate balance, which is a plant decision made every hour at the cement mill. | initial set >30 min |
SO₃ against the optimum for that clinker’s C₃A, and whether the optimum was ever determined or merely inherited. |
| 4. It is sold under a standard the buyer’s specifier already accepts | A cement that cannot be named on a drawing cannot be sold, whatever it does. Standards acceptance, not chemistry, is what gates every low-clinker product. | — | Name the standard and the clause. In India the standards have permitted clinker below 50% since 2015 — so when an Indian plant says it cannot blend further, the constraint is the specifier, the supply of the substitute, or the price, and you should find out which. |
| 5. It reaches the customer inside a freight radius that does not eat the margin | Cement is a low-value, high-mass product. Beyond a few hundred kilometres by road, freight exceeds the ex-works value of the goods, and the business stops being a cement business and becomes a logistics business. | lead distance, km |
Primary lead distance and the freight line in the cost stack. Logistics is the largest single cost line at the largest Indian producer — larger than fuel. Any discussion of plant efficiency that ignores it is discussing the second-biggest number. |
| 6. The buyer cannot easily verify it and does not switch on price alone | This is a market condition, not a product one, and it is the reason the industry is shaped as it is. Cement is 10–16% of concrete by mass but a rounding error on a building’s cost, so the buyer has weak incentive to shop and weak ability to test. | — | Brand premium in the trade segment against the non-trade segment. If the premium is real, the plant is selling reassurance, and an improvement that is invisible to the customer cannot be priced. |
Rows 1 to 3 are made inside the plant, mostly at the cement mill, mostly for free. Row 4 is a standards and market-development problem that takes years. Row 5 is fixed by where the plant was built. Row 6 is the industry’s structure and no plant changes it alone.
The reason this matters for a diagnostic: a plant that is losing on rows 1 to 3 has an operating problem that is cheap to fix and visible to the customer. A plant that is losing on row 5 has a problem no operating improvement will touch. Separating the two is the first judgement of any engagement, and it is made before you look at a single piece of equipment.
§ 05 The quantitative anatomy — what the customer is paying for
Two tables decide how this industry behaves. The first is cement's share of concrete.
| Concrete grade | Cement, kg/m³ | Share by mass | Cost/m³ (RMC, Delhi NCR) | Cement share of cost |
|---|---|---|---|---|
| M20 | ~320 | 13.3% | ₹5,450 | 33% |
| M25 | ~340 | 14.2% | ₹5,900 | 36% |
| M30 | ~380 | 15.8% | ₹6,550 | 37% |
| Lean / mass concrete | ~200–250 | 8–10% | — | — |
That cost table prices cement at the retail bag price of about ₹380 a bag. A ready-mix plant does not pay that — it buys non-trade in bulk. CRISIL puts FY25 non-trade realisation at ₹197–202/bag against a trade MRP of ₹358–363. Re-cost the M25 line at bulk and the cement share of RMC cost falls to roughly 20–22%.
So both figures are true and they describe different customers. For an individual home builder buying bags at MRP and mixing on site, cement is the dominant material cost. For an RMC producer, it is one input among five. Know which customer you are talking about before you quote a share.
Now the consequence, which is the single most important structural fact about this industry. Cement is 10–16% of concrete by mass and 20–37% of concrete cost — and concrete is itself a fraction of the built cost of a structure. A ₹20 per bag price move is about 5% on cement, roughly 1–2% on delivered concrete, and a rounding error on a building.
The customer barely feels a cement price move, so demand is inelastic. But cement is heavy, cheap per tonne and expensive to move, so the market is local — a producer competes seriously only inside a radius of roughly 300–500 km. Inelastic demand inside a freight moat is the definition of a regional oligopoly, and it is why cement behaves like this in every country on earth and never becomes a globally traded commodity like steel.
The second table: what the grades actually promise
Indian standards specify minimum compressive strength at three ages. The chart below is the whole commercial argument between OPC and PPC.
§ 06 What goes wrong here
The customer never complains about kcal/kg. They complain about a slab that would not set, a plaster that bloomed white, a batch that looked different from the last one. Almost every one of those complaints is the visible end of a cost decision taken somewhere else in the plant.
| Field complaint | What the customer sees | Physical cause | The plant decision behind it |
|---|---|---|---|
| False set | Mortar stiffens within minutes, then recovers on remixing with no extra water | Gypsum dehydrated to hemihydrate in a hot mill, then re-precipitating rapidly on mixing | Cement mill outlet temperature too high — above about 110–120 °C. Often the direct result of cutting mill ventilation to save fan power. |
| Flash set | Stiffens irreversibly, with heat. No recovery. | Uncontrolled C₃A hydration — insufficient available sulphate | Under-gypsuming. SO₃ optimised for strength rather than for set; or high-C₃A clinker with gypsum addition left unchanged. |
| Slow or delayed set | Forms cannot be stripped on schedule | Over-sulphation, excess free lime, or very high SCM dosage in cold weather | Over-gypsuming; PPC pushed to the top of the IS range in winter without telling anyone. |
| Low early strength | 3-day and 7-day cubes fail; cycle time slips | Low C₃S, coarse grind, high SCM | High clinker litre-weight chased for fuel economy; separator set for throughput rather than fineness; SCM at the top of the band. |
| Poor 28-day strength | Design strength missed | Free lime, poor burning, inconsistent LSF | Kiln instability and raw-mix control failure — a cross-belt analyser absent or uncalibrated. |
| High water demand | Mason adds water; strength collapses | Particle-size distribution too narrow or too fine | Over-grinding; a high-efficiency separator producing a steep PSD, and the VRM-versus-ball-mill difference never communicated to the customer. |
| Efflorescence | White bloom on plaster and masonry | Soluble alkalis leaching and carbonating at the surface | The alkali cycle in the pyro line — no bypass, high-alkali raw mix, or alkali-bearing alternative fuel. |
| Colour variation | Visible in exposed work | Fe₂O₃ and Mn variation, SCM source change, kiln atmosphere | Raw-mix source switching, or a fly-ash source change mid-campaign. |
Four plant decisions produce nearly all of it: mill outlet temperature, sulphate optimisation, fineness and particle-size distribution, and the alkali cycle. Three of the four are traded against energy cost. That is the thesis of this lesson and the reason it sits at the front of the course — the quality department and the energy department are optimising the same three variables in opposite directions, usually without meeting.
Sourcing note. The mechanisms above are textbook cement chemistry. They are not backed by a published Indian complaint-frequency dataset — none exists. Treat this as a mechanism map, not a ranked frequency table, unless you can get complaint data from the client.
§ 07 Scenarios from the field
The evidence that pricing power is weak at the retail end
On 22 September 2025 GST on cement was cut from 28% to 18%, and India Ratings estimated the pass-through to consumers at ₹30–35 per bag. Retail prices fell by roughly that amount. Producers' realisations also softened afterwards, with commentary pointing at the trade segment in particular.
It is tempting to say "the cut was passed through, so producers have no pricing power," and the course said exactly that in its first draft. The logic does not hold. GST sits outside revenue. A producer's ex-GST realisation is not mechanically reduced by a rate change at all, so any subsequent fall in realisation is a market price movement, not a tax pass-through. The two happened in the same quarter; that does not make one the cause of the other.
What the episode does show, cleanly, is that the ₹30–35 reached the consumer rather than being absorbed into shelf price by the channel — which tells you about retail price formation. If you want evidence on producer pricing power, use realisation against cost across several quarters, not a tax event.
The evidence that pricing power is strong at the brand end
No Indian producer discloses the size of its brand premium. A ₹10–40 per bag "branded versus unbranded" gap circulates widely and could not be sourced — and the comparison may not even exist, since BIS certification is mandatory and effectively all Indian retail cement is branded. What is published: CRISIL's ₹30–60 per bag trade-versus-non-trade price gap, and separately a trade/non-trade net realisation gap of only ₹7–12 a bag. Those two measure different things — list price versus what the producer keeps — and using them interchangeably will not survive a CFO.
So treat the premium in the model below as an input you must source from the client's own price ladder, not as a benchmark. The model shows what a premium is worth if you can hold one; it cannot tell you whether this plant can.
Both things are true at once, and the tension between them is the lesson. Cement cannot hold a tax change at retail, but it can hold a brand premium — because the two are doing different work. Four reasons the brand premium survives:
- The buyer cannot verify the specification. Someone building a house and buying 200 bags cannot test 28-day strength. The brand is a substitute for a test certificate. This is a textbook credence good, and it is why cement is heavily branded in every developing market and largely unbranded in mature bulk markets.
- The decision is made by an intermediary at the point of sale. The mason and the dealer choose the brand far more often than the homeowner does — which is why dealer margin (₹6–8 a bag typically, ₹10–15 where a new entrant is buying shelf space) is a strategic variable rather than a cost line. A widely circulated "70% of decisions are influenced at the point of sale" statistic is attributed to NielsenIQ across the trade press; the underlying study could not be located, so treat the direction as sound and the number as unsourced.
- Failure is catastrophic and delayed. A bad bag shows up as a cracked slab years later, on the buyer's largest lifetime purchase. Risk aversion beats price sensitivity.
- Freight makes the market local. Brand equity only has to be defended against a small competitive set inside 300–500 km. Brand is how a producer holds price inside its own freight moat.
What it is worth, and the comparison that should change how you prioritise
UltraTech runs about 36% of volumes as premium product. A ₹20 per bag premium is ₹400 a tonne on the volume that carries it. Set that against everything the pyroprocessing lesson found: the entire 24-item documented energy register in A7 was worth about ₹129 per tonne of cement, for ₹20.7 crore of capital and a shutdown window.
A ₹20/bag premium held on half a typical trade book — say a third of total volume — is worth roughly ₹134 per tonne of cement across the whole book. That is more than the entire energy register, and it needs no capital, no shutdown and no engineer.
This is not an argument against the register. It is an argument about where a consultant should look first, and about the fact that most cement improvement programmes are staffed entirely by people who cannot touch the larger lever. Work the model in §09 and see it move.
§ 08 The numbers that matter
The Indian product taxonomy, with the strength minimums that define each grade. All strengths in MPa on 70.6 mm mortar cubes per IS 4031 Pt.6.
| Standard | Product | Composition | 3 d | 7 d | 28 d | Fineness min | Your plant |
|---|---|---|---|---|---|---|---|
| IS 269:2015 | OPC 33 | Clinker + 3–5% gypsum | 16 | 22 | 33 | 225 | — |
| IS 269:2015 | OPC 43 | as above | 23 | 33 | 43 | 225 | — |
| IS 269:2015 | OPC 53 | as above | 27 | 37 | 53 | 225 | — |
| IS 1489 Pt.1:2023 | PPC | 15–35% fly ash | 16 | 22 | 33 | 300 | — |
| IS 455:2015 | PSC | 25–70% granulated slag | 16 | 22 | 33 | 225 | — |
| IS 16415 | Composite | Clinker 35–65%, fly ash + slag | 23 | 33 | 43 | 300 | — |
| IS 18189:2023 | LC³ | Clinker 50–80%, calcined clay 10–35%, limestone 5–20% | 23 | 33 | 43 | 400 | — |
| IS 12330 | Sulphate-resisting | C₃A ≤5%; C₄AF+2C₃A ≤25% | 10 | 16 | 33 | 225 | — |
| IS 8041 | Rapid hardening | Clinker + gypsum, no post-burn additions | 27 @72h | — | — | 325 | — |
| IS 12600 | Low heat | Limited C₃S and C₃A | 10 | 16 | 35 | 320 | — |
IS 8112 and IS 12269 no longer exist. The 2015 sixth revision of IS 269 absorbed them, so all three OPC grades now sit under a single specification. Anyone citing "IS 8112 OPC 43" in 2026 is quoting a withdrawn standard — and the QC head will know.
India has no BIS standard for Portland Limestone Cement as a standalone product. Limestone appears only as a ≤5% performance improver in OPC, and as 5–20% inside the LC³ standard. That is a material gap against EN 197-1 CEM II/A-LL and ASTM C595 Type IL, both mainstream in their home markets. Any Indian clinker-factor recommendation must work through PPC, PSC, composite or LC³ — not PLC.
Mapping across markets — and why the mapping is a trap
| India | Europe, nearest | US, nearest |
|---|---|---|
| OPC 33 / 43 / 53 | CEM I 32.5 / 42.5 / 52.5 | C150 Type I, I-II, III |
| PPC (IS 1489 Pt.1:2023) | CEM II/B-V | C595 Type IP |
| PSC (IS 455) | CEM II/B-S or CEM III/A | C595 Type IS |
| Composite (IS 16415) | CEM II/B-M or CEM V/A | C595 Type IT |
| LC³ (IS 18189) | CEM II/C-M (EN 197-5) | C595 Type IT(Q,L) |
| SRC (IS 12330) | EN 197-1 SR cements | C150 Type V |
The strength numbers are not interchangeable, because the test methods are not. IS 4031 uses 70.6 mm cubes of 1:3 cement-to-standard-sand mortar. EN 196-1 uses 40×40×160 mm prisms, broken in flexure then compression, at w/c 0.50. ASTM C109 uses 2-inch cubes at 1:2.75 and w/c 0.485. A cement that tests 53 MPa to IS will not necessarily test 52.5 MPa to EN. Any cross-border specification comparison has to be made on the test method, not the grade label. This is the single most common technical error in cross-market cement consulting.
What actually differs, and what is marketing
| Real, and the customer feels it | Mostly marketing |
|---|---|
| The strength-gain curve, not the 28-day number. The construction cycle is governed by 3-day strength. | Grade above what the design needs. OPC 53 in a 1:6 masonry mortar buys nothing and increases shrinkage and thermal cracking risk. |
| Water demand and workability at a given fineness — drives placeability and pumping cost. | "Higher grade = stronger house." Concrete grade is set by mix design; cement grade is one input among several. |
| Consistency batch to batch. Worth more to an RMC producer than any average property. | Branded fineness claims. Blaine above the minimum is a genuine lever, but it is not disclosed — so it cannot be a purchase criterion. |
| Durability in the actual exposure — sulphate, chloride, alkali-silica. Real, but only where the exposure exists. | "Special" product naming where the underlying product is the same PPC to the same IS 1489 Pt.1. The differentiation is packaging and channel. |
| Heat in mass pours. Only IS 12600 specifies it: ≤272 kJ/kg at 7 days. |
The consultant's line: the customer is buying a strength-gain schedule and a consistency guarantee. The grade label is a proxy for the first and says nothing at all about the second.
§ 09 Build the model
First, by hand
Step 1. A bag is 50 kg, so there are 20 bags in a tonne.
Step 2. ₹1 per bag is therefore ₹20 per tonne of cement — on the volume that actually carries the premium.
Step 3. A 5 Mt/yr producer with a 67% trade book, holding a ₹20/bag premium on half of that trade volume, carries it on 5 × 0.67 × 0.5 = 1.675 Mt.
Step 4. ₹400/t × 1.675 × 10⁶ t = ₹670 × 10⁶ = ₹67 crore a year.
Now normalise it so it can be compared with anything else in the course: ₹67 crore over the full 5 Mt book is ₹134 per tonne of cement. Hold that number.
Then, with the model
The comparison bar is fixed at the A7 register's value — about ₹129 per tonne of cement, for ₹20.7 crore of capital and a shutdown window. Move the sliders and find the premium and capture rate at which the commercial lever beats the entire engineering programme.
The premium is not free, it is just not capital. Holding it needs brand spend, distribution depth and dealer economics — an operating expense rather than a capex line, which is exactly why it is easier to approve and easier to cut. On the size of that spend, be careful: a claim that Indian producers spend 0.2–0.5% of revenue on brand against 0.7–1.0% for global majors is unsourced and appears to be wrong. The one hard Indian figure available is that thirteen listed Indian cement companies spent ₹3,076 crore on advertising and sales promotion in FY23, UltraTech alone ₹1,296 crore, at up to about 2% of revenue — a combined line, so not strictly comparable, but an order of magnitude above the claim.
A premium can be competed away; a kcal cannot. The energy register is a permanent change to the plant's cost base. A price premium is a position that has to be defended every quarter, and the GST evidence above shows how quickly the retail end gives ground.
The capture rate is the whole model and it is the number you will not be given. "Half the trade book at ₹20" is an assumption, not a finding. Ask for the price ladder by SKU and district before you use any of this in front of a client.
§ 10 Upgrade paths — changing what you sell
Product-mix change follows the same tiering logic as plant upgrades: sort by what it preserves, not by what it costs.
| Tier | What it needs | Interventions | Capex | Shutdown |
|---|---|---|---|---|
| 0 · Commercial | Nothing in the plant | Brand premium, channel mix, dealer economics, SKU price ladder, premium-product positioning | opex only | none |
| 1 · Mix shift | Existing capability | Move the OPC/PPC split within the SCM band already permitted; re-optimise Blaine and PSD per product | nil | none |
| 2 · Bolt-on | Existing mills and silos | Storage and dosing for a second SCM; a dedicated silo for a premium product; packer and bag changes | ₹5–40 cr | days |
| 3 · New capability | Existing clinker line | Composite cement capability; a calcined-clay line for LC³ — the only SCM that is not another industry's by-product | ₹100 cr+ | weeks |
Fly ash depends on coal-fired power. Slag depends on blast-furnace ironmaking. Both are shrinking assets outside a cement producer's control, and neither can be secured by contract indefinitely. Calcined clay is a mineral you can lease. India has the standard (IS 18189:2023), abundant kaolinitic clay, and commercial production already running at JK Lakshmi's Jaykaypuram plant. That combination — standard, resource, and supply-chain independence — is why LC³ deserves more attention than its current near-zero market share suggests.
§ 11 The frontier, and who is running it
The frontier in what is actually sold is low-clinker cement, and it is moving faster in the standards than in the plants. The standards in Europe, the United States and India already permit clinker factors far below what anyone ships. The binding constraint is specifier acceptance and substitute supply, not the specification.
| Status | What | Plant, company, country | The number |
|---|---|---|---|
| OPERATING | The oldest and largest commercial calcined clay line | Rioclaro — Cementos Argos, Colombia. “Cemento Verde”. Production started February 2020. | 450,000 t/yr calcined clay, US$78m. Claims up to 38% lower CO₂ and up to 30% lower energy. Clinker factor not disclosed. It predates every “world first” claim in this field by three years. |
| OPERATING | Flash calcination at scale | Kribi — Cimpor, Cameroon. Flash calciner line launched October 2023, plant inaugurated July 2024. | Claimed 0.4 Mt/yr calcined clay on a 1.2 Mt/yr integrated plant. A trade-press technical review lists the same line at 720 t/d, which is about 0.24 Mt/yr. One of those is design and one is achieved, and no source says which. |
| OPERATING | A kiln converted rather than built | Guayaquil — Holcim, Ecuador. Existing clinker kiln retrofitted to calcined clay, commissioned first half of 2025. | “Up to 2 million tonnes of calcined-clay-based cement per year”. Read the denominator — that is cement, not clay, and the two differ by roughly the substitution rate. |
| OPERATING | Europe’s newest calcined clay line | Čížkovice — Holcim Česko, Czech Republic. Commissioned September 2026. | 115,000 t/yr calcined clay, enabling about 580,000 t/yr of the low-clinker product. CZK 1bn, of which CZK 330m is a Modernisation Fund grant. Divide the capex by the clay and you get about US$415 per annual tonne; divide it by the cement it enables and you get about US$82. The denominator moves the answer five-fold, and both versions are in circulation. |
| OPERATING | India, commercial LC3 | JK Cement (“JK Super LC-3”, October 2025) and JK Lakshmi Cement, Jaykaypuram, Sirohi, Rajasthan (“Green PRO LC3”, first dispatches February 2026). | Both claim about 40% lower CO₂. Neither discloses a capacity, a clinker factor, or in JK Lakshmi’s case the standard. JK Cement’s product is sold under IS 18189:2023. The marketing is well ahead of the disclosure. |
| STANDARDS | What is already permitted | — | EN 197-5:2021 created CEM II/C-M at 50–64% clinker and CEM VI. EN 197-6:2023 admits recycled building materials. ASTM C595 Type IL (limestone, since 2012) and Type IT (ternary, since 2009). India: IS 16415:2015 composite cement already permits clinker at 45%, and IS 18189:2023 covers calcined clay limestone cement at 50–80% clinker. India’s standards stopped being the constraint in 2015. |
| THE DATA | What the customer says about low-clinker cement | Joint ACI–ASCC survey on Type IL concrete, United States, August–September 2023. 173 respondents, mostly ready-mix producers, contractors and designers. | 56% reported problems more often with Type IL than with ordinary Portland. Admixture suppliers 89%, owners 80%, concrete contractors 72% — but 86% of cement producers said “same frequency”. Issues: saw-cut timing ~70%, water demand 60–75%, finishing ~45%, setting time ~45%, shrinkage cracking ~45%. The split between who makes it and who places it is the whole finding. |
| FAILS CHECKING | “World’s largest calcined clay plant” | Tema — CBI Ghana with Heidelberg Materials. | Government material says 1.5 Mt/yr and US$110m; the company’s own release says more than 400,000 t/yr of calcined clay and claims only the largest flash calciner. Four capex figures circulate (US$66.7m with a US$20m loan, US$100m, US$110m) and three commissioning dates. Argos Rioclaro has run 450,000 t/yr since 2020. Largest flash calciner, plausibly. Largest calcined clay plant, no. |
World calcined clay output in 2023 was 2.1 Mt — 0.05% of global cement production. Forecasts for 2035 range from about 20 Mt (industry analysis) to far higher in policy scenarios, and the long-run forecasts diverge by a factor of three between the IEA and independent analysts. A September 2026 project-pipeline count found 14 projects worth US$405m and 3.3 Mt/yr, of which five were delayed and five had no commissioning date.
So: a real technology, commercially proven at several named plants, and currently a rounding error on world production. Both halves belong in the same sentence when a client asks about it.
Disclosed company averages: Dalmia Bharat 59.7% (FY25), the lowest credible figure from a large producer; UltraTech 67.6% (Q2 FY26); Indian industry 68.0% (FY25), up from 70.4% in FY20. EU about 78%, world about 82% on the clinker-to-cement basis. The United States: 69 Mt clinker against 84 Mt cement in 2025, about 82%, with blended cement at 63% of shipments.
No clinker factor has ever been published for a single named plant. Every figure above is a company or national average. And the two common denominators — clinker-to-cement and clinker-to-cementitious — are not the same measurement, which is why the same industry is described as being at 82% and at 72% in documents published the same year. Reconcile the denominator before comparing any two of these.
§ 12 What's changing now
Current as of August 2026 · refresh every six months"GST 2.0." India Ratings estimated the pass-through at ₹30–35 per bag. UltraTech's Q3 FY26 numbers showed a 3% sequential fall in blended realisation, concentrated in trade.
What it means: the cut reached the consumer rather than the producer. Any 2026 price analysis that does not adjust for this will read a tax change as a demand signal.73% of Indian production in 2023, 76% in FY25, 76–78% expected FY26. LC³ is now commercially produced at JK Lakshmi under IS 18189:2023, and Noida International Airport is the first large-scale LC³ project.
What it means: at 1.5 points a year the remaining 24% of OPC takes sixteen years to displace — and it will never fully displace, because OPC serves the 3-day-strength requirement that PPC cannot. RMI's 2026 study puts the LC³ blockers on the concrete side: no standard mix designs, ready-mix plants not operationally ready, and no EPDs.The definitive regime began, with cement covered. The default emission value for "other countries" is 1.551 t CO₂e per tonne of clinker; Turkish producers' actual reported emissions are around 0.88. That is roughly €20/t against €80/t of carbon cost.
What it means: the penalty is not the carbon price, it is the gap between the default and your verified actual — about €60 a tonne of clinker. And the verification window cannot open until 1 January 2027 and must close by September. The constraint is calendar, not capital, exactly as it is with plant upgrades.§ 13 Check yourself
§ 14 Mini case
| Parameter | Quarter start | Quarter end | Change |
|---|---|---|---|
| Cement mill outlet temperature, °C | 98 | 124 | +26 |
| Mill vent fan power, kWh/t cement | 6.4 | 5.6 | −0.8 |
| Gypsum addition, % | 4.2 | 4.2 | 0 |
| SO₃ in cement, % | 2.10 | 2.08 | −0.02 |
| Blaine, m²/kg | 312 | 318 | +6 |
| Fly ash, % | 28 | 29 | +1 |
| Clinker C₃A, % | 7.8 | 7.9 | +0.1 |
| 28-day strength, MPa | 41 | 40 | −1 |
| Initial setting time, minutes | 145 | 140 | −5 |
What is happening, and what caused it?
What the data says
This is false set, and the tell is the description, not the test. Stiffening within minutes that recovers on remixing without extra water is the textbook signature. Flash set does not recover and evolves heat; this did neither. And critically, false set does not show up in any of the tests the quality manager ran — strength, fineness and setting time all passed, because the standard setting-time test remixes the paste.
The cause is the mill outlet temperature. 98 → 124 °C takes the mill above the point where gypsum (CaSO₄·2H₂O) dehydrates to hemihydrate. On mixing, hemihydrate re-precipitates as secondary gypsum within minutes and stiffens the paste — then breaks down again under remixing. Everything else is stable: gypsum addition unchanged, SO₃ unchanged, C₃A unchanged, so this is not under-gypsuming.
And look at why the mill got hot. Vent fan power fell 0.8 kWh/t in the same quarter. Someone reduced mill ventilation — almost certainly as an energy-saving measure, and quite possibly logged as one.
Now size both sides of it
| Side | Working | ₹ crore / yr |
|---|---|---|
| The saving that caused it | 0.8 kWh/t × 2 Mt × ₹4.73/kWh | 0.76 |
| The volume it cost | District is 4% of a 1.5 Mt trade book = 60,000 t a year; 18% down = 10,800 t/yr at ₹1,000/t margin | 1.08 |
| The premium at risk | Brand damage across the district's dealer network — unquantified, and the reason the number above is a floor | — |
| Net | A ₹0.76 crore saving that has already cost more than it saved | −0.32 |
The fix is to restore mill ventilation and add water spray for temperature control, or move part of the gypsum to a form that resists dehydration. Neither is expensive. The expensive part has already happened.
₹1.08 crore lost against ₹0.76 crore saved is a net of about ₹32 lakh — small enough that no one will ever go looking for it. That is the point. If the saving had been wiped out eight times over, someone would have investigated. A loss that is merely slightly larger than the gain is invisible, because the gain is booked in an energy report with a name on it and the loss is spread across forty dealers with no name on it at all.
And the ₹1.08 crore is a floor: it counts only volume already lost in one district, not the brand effect, not the other districts, and not the cost of winning the dealers back. Size the visible half, then say plainly that the invisible half is larger.
Nobody did anything careless. The energy engineer found 0.8 kWh/t and booked it. The quality manager ran every test in the standard and they all passed. The failure was that no test in IS 1489 detects false set, and no report links mill ventilation to a dealer complaint.
This is the pattern the whole lesson exists to teach: the quality complaint is the visible end of a cost decision taken elsewhere in the plant, by someone who will never hear about it. When you are handed an energy-saving register — like the 24 items in A7 — the right question is not only "is it real?" but "what does it touch downstream?"
§ 15 Go deeper
§ 16 Carry forward
- "Cement is about 14% of concrete by mass but a third of its cost at bag prices and about a fifth at bulk. Which of those is your customer? Because the answer changes what your price actually means to them."
- "Your customers aren't buying 53 MPa at 28 days. They're buying 27 MPa at three days, because that's what lets them strip formwork — that's why OPC survives at a quarter of the market despite costing you more to make."
- "Almost every quality complaint you get traces to four decisions: mill outlet temperature, sulphate optimisation, fineness, and the alkali cycle. Three of those are being traded against energy cost by people who never see the complaint."
- "A ₹20 a bag premium on half your trade book is worth more than every energy project in your plant, and it needs no capital and no shutdown. Can I see your price ladder by district before we talk about the kiln?"