Cement Is 8% of Global CO2. What Can a House in Nepal Actually Use Instead?

If you are building a house in Nepal right now, ordinary Portland cement (OPC) is still the default. It is cheap, every mason knows it, you can buy it in every bazaar from Mechi to Mahakali, and it passes the building code without an argument.

It is also the most carbon-intensive thing you will ever buy. Cement alone is roughly 7–8% of global CO2 — about 2.5–4 Gt a year depending on how you count energy. If cement were a country it would be the third-largest emitter, after China and the US. So the question is not "is OPC bad?" We know it is. The question is: what can you actually use instead, in Nepal, without the house cracking, leaking, or bankrupting you?

That is what this post is about. Plain language first, engineer details where they matter, and a comparison table measured the same way: 1 m² of wall, same structure and comfort, over 50 years.

Why cement eats the climate: you are un-baking limestone

Limestone exists because ancient seas pulled CO2 out of the air. CO2 dissolved as weak acid rain, reacted with rocks, washed to the ocean, and settled as calcium carbonate (CaCO3) — a giant, million-year carbon store.

Making cement runs that film backwards:

CaCO3 + heat (~1,450 °C) → CaO + CO2

About 60% of cement's CO2 is this chemistry itself — unavoidable as long as you make clinker. You cannot efficiency your way out of it. The other ~40% is the fuel to reach that heat, plus grinding and transport. A tonne of clinker releases roughly 0.85–0.95 tonnes of CO2.

That is why OPC is so hard to kill. Limestone is everywhere, the kiln process is a century optimized, and the product sets in hours and lasts decades. Every alternative has to beat that triangle: price, reliability, scale. Most beat one corner and lose the other two.

To cut emissions you only have four real levers:

1. Less clinker per bag — replace clinker with calcined clay, limestone, fly ash, slag, ground glass (LC3 does this; ~50% less clinker, ~30–40% less CO2 in literature).
2. Less cement per wall — stabilized earth (CSEB uses 5–10% stabilizer instead of a full cement block), better design, build once and maintain for 50 years.
3. Cleaner heat — alternative fuels, efficient kilns. Helps the 40%, does nothing for the 60%.
4. Let it reabsorb some CO2 — concrete slowly carbonates over decades. Real, but small and slow; never the plan.

Everything below is one of those four, or it is a distraction.

Quick comparison: same wall, same 50 years

Sort any column, click a row for cost, carbon, water, and what it would take to scale it in Nepal. Strengths are literature bands, not Nepal lab guarantees.

7-day is handling, by 28 days it's settled

Engineers quote two numbers because they answer two different site questions.

A normal cement block hits ~65% of its strength at 7 days (say 5–13 MPa) and full strength at 28 days (7–20 MPa). At 7 days you decide: can I demould it, stack it, carry it? At 28 days you decide: can this wall carry the house?

That trajectory calibrates the whole table. CSEB sits at 2–5 MPa at 7 days — do not load it, do not rush it, cure it 28 days. LC3 behaves almost like OPC (8–20 MPa at 7 days). Pure fly-ash geopolymer looks dead at 7 days unless you heat-cure it or blend slag into it. The rule for Nepal: 10 MPa at 28 days is the load-bearing threshold for a low-rise wall. Below that is partition, infill, or insulation — not structure. And "child-safe" is a separate toxicity question (leachate, dust, alkalinity), never infer it from MPa.

What material science actually offers (and what is hype)

LC3 — the only national-scale bet. Limestone + calcined clay + 50% clinker. Clay calcines at ~750 °C, half the heat of clinker, and both raw materials exist in Nepal. Mason-familiar, normal water curing. What is missing is a reactive-clay survey and calcination capacity. If I had to pick one industrial investment for Nepal, this is it. Cost: Low. TRL 7.

CSEB — the other scalable one. Soil you already have, pressed on site, 5–10% cement or lime. Carbon is tiny because binder is tiny and transport is near zero. Build Up Nepal has already shown it works here. The catch is discipline: soil selection, presses, trained masons, 28-day cure, vertical reinforcement in seismic zones, and saturated strength runs ~35% below dry. Slow early strength means a bigger curing yard. Cost: Low. TRL 8.

Geopolymers (fly ash + slag, bagasse ash, recycled fines) — pilots, not replacements. The chemistry works: 10–30 MPa at 7 days with slag, ambient cure, no kiln. But Nepal has no domestic coal-ash or blast-furnace-slag pipeline — both come from India — and the activator (Na2CO3 cheaper, NaOH/silicate worse) is imported and dominates both cost and CO2. Bagasse ash only exists seasonally near Terai mills. Recycled fines from Kathmandu demolition vary stream by stream and need per-batch QA plus leachate testing. Cement-free does not mean low-carbon until the activator dose is fixed and counted. Keep these as regional pilots.

Magnesium phosphate, fiberboard, AAC — category errors if sold as "cement replacements." Mg-phosphate hits 15–35 MPa in 1–3 days and is brilliant for rapid winter repair in dry cold places — then water reverses it. Fiberboard (2–8 MPa bending) is an interior/insulation panel, not a block. AAC (3–7 MPa) is factory-steam infill that still contains cement, needs plaster and seismic detailing, and is water- and energy-hungry to make. Useful products, wrong job description.

Engineer checklist — demand this before any "low-carbon" claim

• 7-day + 28-day strength at stated curing
• Activator dose as Na2O-eq
• LCA boundary (cradle-to-gate at minimum, 1 m² wall over 50 years preferred)
• Water regime
• Durability (wet-dry, freeze-thaw, immersion) — never compression alone

For modelling, thermodynamic screening (CemGEMS) and mix optimization (SLAMD) help, but only after XRF/XRD/PSD of your local inputs.

The part nobody prices: heat

Kathmandu and the Terai are getting hotter, and concrete roofs and dark cement plaster make it worse. Two things matter: reflectivity (light surfaces bounce sun back) and thermal mass + insulation (heavy earth walls flatten day-night swings; insulation keeps heat out of the roof).

Practical, Nepal-realistic moves:

• Light-colored lime or white-cement finish over dark grey plaster cuts roof and wall surface temperature several degrees. Cheap, proven.
• Exposed CSEB / earth walls with a good overhang outperform thin cement-block walls for comfort in the mid-hills — heavy, breathable, no AC needed most of the year.
• Roof is the heat inlet. Insulation panels (even agro-fiber boards used as insulation, not structure) plus a reflective top coat beat a bare concrete slab.
• Terai lowland: prioritize mould/ASR resistance and ventilation over mass. High hills: prioritize freeze-thaw and keeping Mg/phosphate and earth dry in service.

No binder in the table "solves" urban heat island by itself. But LC3/CSEB walls + light finish + shaded roof is measurably cooler than a dark OPC box, at the same structural target.

What makes sense in Nepal, by place

• Mid-hills (Kathmandu, Pokhara): LC3 where clay + kiln exist; CSEB where presses + trained masons + NBC acceptance exist. Recycled-fines pilots near Valley demolition streams.
• Terai: bagasse-ash geopolymer near sugar mills with sunlight cure; LC3/CSEB otherwise. Design for hot-humid: mould, moisture, flooding.
• High hills + Himalaya: keep it simple and dry. OPC/LC3 + stone, Mg-phosphate only for rapid cold repair, earth/fiber only where they stay dry. Freeze-thaw testing is non-negotiable.

Across all three, the binding limits are rarely water. They are: reactive-clay survey, ash/slag + activator imports, presses + masons, glass collection + grinding, CDW sorting + QA. Whoever solves those supply chains decides what scales.

Building a home now? Use this decision

Default: OPC is still the most economical and lowest-risk material for a single new home. Say that plainly. The supply chain exists, the mason exists, the engineer will sign it, the rate is known.

Deviate from it for a reason:

• Pick CSEB if you have suitable soil on or near site, can hire a trained crew, have room to cure 28 days, and will do seismic reinforcement properly. You save carbon and often cost on the wall, but you pay in supervision and time.
• Pick LC3 the moment it is reliably available from a nearby plant with test data — same crew, same curing, ~a third less CO2, little behavior change. Ask the supplier for 7/28-day data and clay reactivity (R³), not slogans.
• Pick AAC only as lightweight infill in a framed structure with proper plaster and seismic detailing — never as cheap "load-bearing" to save cement.
• Do not pick geopolymer, Mg-phosphate, or fiberboard for your whole house today unless you are running a supervised pilot with an engineer and lab backup. The activator, water-sensitivity, and QA risks are yours alone.

And whatever you pick: cure it (water costs less than a cracked wall), plaster and maintain it (a 50-year wall beats two 25-year walls on carbon every time), keep dust and fresh-alkali exposure controlled on site, and test red mud / fly ash sources for leachate before bringing them near kids or wells.

Takeaways

1. Cement is ~8% of global CO2 because making clinker un-bakes limestone. Only less clinker, less cement per wall, cleaner heat, or longer life moves the needle.
2. For Nepal at scale: LC3 and CSEB. Everything else is a regional pilot or a different job (repair, infill, insulation).
3. 7-day is handling, by 28 days it's settled — 10 MPa at 28 days is the load-bearing line for low-rise walls.
4. If you build now, OPC is still the safe default — beat it with CSEB for walls where crew + cure exist, LC3 where supply exists, and light finishes + roof insulation for heat regardless of binder.