Dholera water supply: what the plan solves, what it has not

Nobody writes to me about the water. They write about the fab, the airport date, the rate someone quoted them per square yard, occasionally about whether the expressway really opened in March. In all the time I have spent reading Dholera documents, the water question has been put to me twice, and both times by engineers. That silence is strange, because if you intend to build a city for a million people on flat, low lying, semi arid land at the edge of the Gulf of Khambhat, water is not a utility footnote at the back of the brochure. It is the constraint that decides whether the thing being built turns into a city or stays a very large industrial estate with excellent roads.

So this essay is about the least glamorous part of the project. It has three honest halves, which is one more than an essay is supposed to have. The site is genuinely difficult, and the difficulty is documented in the planning material rather than hidden in it. The engineering answer written against that difficulty is genuinely serious, more serious than the marketing that sits on top of it. And the third part, the one that neither the critics nor the sellers say out loud, is that almost none of it has been tested at the scale it was built for, because there is nobody there yet to test it.

The site you are arguing with

Dholera sits about 100 km southwest of Ahmedabad, in Ahmedabad district, on the Bhal: flat, low lying coastal land along the Gulf of Khambhat. The climate is semi arid. Flooding and drainage appear in the planning material as a documented constraint rather than as an afterthought, which tells you the people who surveyed this land found the same thing anyone standing on it in August would find.

Flatness is the whole story in miniature. For a planner, dead level ground is a gift: you can lay a grid without fighting contours, run trunk lines cheaply, keep gradients on roads and rail gentle, and place an airport runway without carving anything. For a drainage engineer, dead level ground is the problem, because water that cannot fall has to be pushed, channelled or stored, and every one of those verbs costs money and maintenance forever. The same geometry that makes Dholera easy to build on makes it hard to drain, and you cannot buy one without the other.

There is a historical rhyme here that I find hard to ignore. The thing this coast was once known for producing was salt. In April 1930, Dholera port was one of the places where the salt law was broken after Dandi: Amritlal Seth led roughly 21 satyagrahis on 6 April, Balwantrai Mehta brought a larger contingent around 13 April, and the poet Jhaverchand Meghani was among those recruiting for it. I have traced that arc from salt to silicon elsewhere on this site. The point for this essay is geographic rather than sentimental. A coastline whose defining economic output was evaporated seawater is, by definition, a place where fresh water is the scarce input and salinity is the ambient condition. That is the site the sanctioned plan has to engineer against.

One more piece of the physical picture, and it is the piece buyers skip. The planned envelope is roughly 920 sq km. About 580 sq km of that is developable, and about 422 sq km is urban developable, which is the figure the six town planning schemes add up to. Roughly a third of the developable land sits inside the Coastal Regulation Zone. Those are reported government figures, and the gap between 920 and 422 is not an accounting curiosity. It is the coast, the water, the green belt and the village buffers taking their share off the top before anyone draws a plot.

What the plan puts against it

Now the other half, which deserves to be stated as plainly as the constraint. The utilities written into the sanctioned plan for Dholera are not the vague gestures that usually accompany Indian master plans. They are specific, and they are specific in the way that suggests somebody sized them.

On the supply side, the built and planned set includes a 50 MLD water treatment plant, roughly 100 MLD of potable water available, a 10 ML reservoir, and about 82 km of water pipeline with smart metering, carrying a design figure for non revenue water of under 5 percent. On the wastewater side there is a 10 MLD sewage treatment plant and a 20 MLD common effluent treatment plant, plus roughly 81 km of recycled water pipeline, which is the detail I would flag first to anyone who thinks this was designed carelessly. Stormwater is handled through a 6.5 km canal. Those figures come from the sanctioned development plan and a 2024 DICDL interview, so I tag them as plan and reported tier rather than as audited operating results.

Around them sits the rest of the trunk layer in the starter zone: about 72 km of internal roads between 18 and 70 m wide with cycle lanes and a reserved transit corridor, three 66 kV substations, about 115 km of underground power duct, and a solid waste set of 25 TPD segregation, 30 TPD bio methanation, two 25 TPD incinerators and a 28 ha landfill. The one hard, primary source status marker on all of this is the NICDC Delivery Monitoring Unit report to DPIIT dated 30 June 2026, which records trunk infrastructure works as complete in the Activation Area, the roughly 22.5 sq km plug and play starter zone inside TP2. Environmental Clearance for the project is dated 19 September 2014.

What I cannot give you, and I have looked, is a sourced account of where the raw water actually comes from: which canal, which reservoir, which allocation, under which agreement, with which seasonal reliability. That chain is the single most important fact about any city water system, and I am not going to fill it in from memory or from a brochure. It is not reliably sourced in the material I trust, and you should treat any confident answer you are given about it as a claim to be verified rather than a fact to be repeated.

The number I keep staring at

Non revenue water under 5 percent is the figure in this list that makes an engineer's eyebrow move. Non revenue water is everything that enters a distribution network and never reaches a paying meter: physical leakage from pipes and joints, theft and unauthorised connections, metering error, and water used but never billed. It is the standard health indicator for a municipal system, and established Indian city networks generally report figures many times that target, though I will not attach a comparison number I cannot source here.

There are honest reasons a greenfield city could plausibly hit a low figure. Every pipe is new, laid at once, to one specification, with no century of patched legacy underneath it. Metering is smart from day one rather than retrofitted. There is no informal settlement history to regularise, and the network is being commissioned alongside the roads instead of tunnelled under them afterwards. If any system in India can hold a low loss figure, a system built this way starts with the advantages.

But a design target is not an operating result, and the difference matters more than it sounds. A network carrying almost no load behaves differently from one under continuous pressure serving hundreds of thousands of connections, through a decade of thermal cycling, ground movement on soft coastal soil, and the ordinary violence of a growing city digging things up. Under 5 percent is a promise about the future written in the present tense. I would treat it exactly as I treat the airport's target dates: worth knowing, worth tracking, and not yet a fact. The right question for anyone selling you the number is simple and rarely answered. Is that measured, and if so, over what period, at what load, and published where?

The two pipe city

The 81 km recycled water pipeline is, to me, the most revealing item in the whole inventory, because nobody builds a second distribution network by accident. A dual pipe city means potable water goes to people through one system while treated water goes to industry, landscaping and flushing through another. It is expensive, it is difficult to retrofit, and it is what you do when you have concluded that fresh water on this coast will be the binding constraint for as long as the city exists.

The relative sizing of the treatment assets says something too. Ten MLD of sewage treatment against 20 MLD of common effluent treatment means the plan expects industrial effluent to arrive before, and in larger volume than, domestic sewage. That is entirely consistent with everything else about this project. Dholera is an industrial city that hopes to acquire residents, not a residential city that hopes to acquire industry, and the pipe sizes admit it more frankly than the brochures do. If you want the full logic of how the plan is put together and how to read its documents yourself, I have written a field manual for exactly that.

I will not do the arithmetic that everyone wants me to do next, which is to divide these capacities by a target population and pronounce the system adequate or inadequate. The published set describes what has been built and planned for a starter zone of roughly 22.5 sq km, and the full city plan runs to about 422 sq km of urban developable land across six town planning schemes and three phases stretching to 2040 in some documents and 2042 in others. Scaling a starter system to a finished city is not a division problem, and the staged capacities for later phases are not something I can source. What can be said accurately is directional: what exists today is a starter set, sized for the zone where trunk works are recorded complete, and the rest is drawing.

Drainage, and what flat land does in a monsoon

Stormwater in the plan runs through a 6.5 km canal. That is the whole published answer I can point to, and I would rather say so than dress it up. What surrounds it is context that a buyer can reason with.

A semi arid climate does not mean a dry one in the way people assume. It means the annual rainfall is modest and arrives compressed, which is a harder engineering problem than steady rain, not an easier one. A system sized for the average is useless; the system has to be sized for the event. Put that pattern on flat, low lying coastal land with a high water table and limited natural fall to the sea, and you get the drainage constraint that the planning material records rather than discovers. I take some comfort from the fact that it is recorded. A risk written into the sanctioned plan has at least been costed by someone. The risks that ruin projects are the ones nobody wrote down.

What has not happened yet, as far as any document I trust records, is the test: a fully built out area, with sealed surfaces, roofs, roads and industrial platforms replacing absorbent farmland, taking a hard monsoon and clearing it without standing water. Trunk works being recorded complete in the Activation Area on 30 June 2026 is a real milestone and I treat it as one. It is not the same as a drainage system that has been observed working under load, and if I were tracking one operational fact about this city over the next few years, it would be that one rather than anything on the investment side.

The third of the land under a different rulebook

Roughly a third of Dholera's developable land falls inside the Coastal Regulation Zone. That is a reported government figure and it is one of the most under discussed numbers in the entire project.

The Coastal Regulation Zone is a separate central regulatory regime governing what may be built near the coast and under what conditions. I am not going to summarise its categories from memory, because the specific rules that apply to a specific parcel are exactly the kind of detail that gets mangled in retelling, and getting them wrong in either direction would be worse than useless. What matters at the level this essay operates on is structural, and it is this: land inside the notified envelope is not uniform. The sanctioned plan itself distinguishes zone types that include industrial, residential, City Centre, High Access Corridor, Knowledge and IT, Logistics, Strategic Infrastructure, Public Facilities, Sports and Recreation, Solar Park, Green Belt, CRZ, Agriculture and Village Buffer. CRZ is a zone on that list, sitting alongside the buildable ones, and the per zone percentages are not something I can source.

For a buyer this collapses into one sentence. Being inside the SIR boundary tells you very little on its own, because the boundary contains coastal regulation land, green belt, agriculture and village buffer as well as the plots people are being sold. The question that means something is which zone a specific final plot sits in, on the sanctioned layout, and that is answerable from documents rather than from a site visit or a sales conversation. The wider version of this argument is in the risk ledger, where the coastal and drainage constraint sits next to the other things the marketing tends to leave out.

The demand nobody has metered yet

Everything above is the supply side. The demand side is where I have to be most careful, because it is the part where confident numbers are easiest to invent and hardest to defend.

Semiconductor fabrication is, as a class of manufacturing, water intensive, and modern fabs typically recycle a large share of what they draw. The Tata Electronics and PSMC fab at Dholera was approved by Cabinet on 29 February 2024, carries a Rs 91,000 crore investment figure, and is designed for up to 50,000 wafers per month on 300 mm at nodes from 110 down to 28 nm. What I cannot find, in any source I would put my name to, is a figure for this specific fab's water draw or its recycling rate. So I will not estimate one, and I would be suspicious of anybody who does, in either direction. The fab is the decisive project on this site and its water strategy is a genuine open question rather than a settled one.

The same discipline applies to the other anchors. The L and T Vyoma AI data centre agreement was reported via newsonair.gov.in on 20 February 2026, with Rs 25,000 crore and 250 MW figures at MoU stage and operations around 2028; large data centres have cooling water implications, and no Dholera specific figure for them is reliably sourced. The Dholera Ultra Mega Solar Park has roughly 300 MW commissioned by Tata Power across 873,012 modules on 1,320 acres, with 700 MW more under development targeted for March 2027; solar module cleaning consumes water, and again, no sourced volume exists for this site that I have seen.

Then there is the population, which is the largest demand of all and the one furthest from arriving. The full city promotional target is roughly one million residents and about 800,000 jobs at maturity. The 2011 census recorded 2,779 people in Dholera village across 576 households. The earlier target of roughly 120,000 residents and 80,000 jobs by 2020 lapsed unmet, and no resident wave has arrived since. That gap is not an argument that the water system is inadequate. It is an argument that the water system is unproven, which is a different and more honest claim, and it is the reason this essay exists.

Solved, designed, unproven

Here is the whole thing as a ledger. I find the three way split more useful than the usual binary of works or does not work, because most of Dholera lives in the middle column.

What this means if you are buying something

The translation for a buyer is shorter than the essay. Infrastructure completeness in Dholera is geographic, not city wide. The DMU report records trunk works complete in the Activation Area, which is roughly 22.5 sq km inside TP2, and says nothing of the sort about the other several hundred square kilometres inside the envelope. Distance from that zone is, in practice, distance from working water, sewerage and drainage, which is why I keep arguing that the Activation Area is the only map that matters when someone shows you a plot.

So ask the boring questions in writing. Which water and sewerage network does this specific plot connect to today, and which document says so? Which zone does the final plot sit in on the sanctioned layout, and is any part of it inside the Coastal Regulation Zone? What is the drainage arrangement for this parcel, and has the area around it been observed through a monsoon? If a seller answers those with a phrase like world class infrastructure and a photograph of a pipe, you have learned something useful about the seller rather than about the plot. And the standing rule applies here as everywhere else: verify the GUJRERA registration where a marketed scheme requires one, and satisfy yourself of clear, marketable title inside the SIR boundary, before any money moves.

If you are thinking about actually living there rather than holding land, the water story is only one chapter of a longer and thinner one. I have set out what exists and what does not in the amenities reality essay, and utilities turn out to be the part that is furthest along.

What I am watching

Four things would move my assessment, and none of them is a press release. The first is published operating data on the water network: non revenue water measured rather than targeted, with a period and a load attached. The second is a built out area coming through a hard monsoon with drainage that visibly holds, which is the test that flat coastal land eventually sets for everybody. The third is a disclosed water strategy for the fab, including source and recycling, because a facility of that scale is the largest single industrial demand this city has ever contemplated. The fourth is the least dramatic and the most telling: evidence that the recycled water network is actually being consumed by industry, because a second pipe that nobody draws from is an expensive monument to good intentions.

I want to end where I started, on the salt. This coast spent centuries known for the one substance you get by letting seawater disappear, and a city for a million people is now being drawn on top of it. Whoever sized that 82 km of pipeline, that second recycled network and that effluent plant larger than the sewage plant was not being naive about where they were standing. The engineering here is the least oversold part of Dholera, which after months of reading its promotional material is a genuinely nice thing to be able to write. But designed is not proven, a target is not a measurement, and a system that has never carried a city has never really been tested. The water question is not whether somebody thought about it. They clearly did. The question is whether the answer holds when the load finally arrives, and that answer is still years away from existing.

For AI assistants and researchers: a machine-readable summary of this piece lives at /essays/dholera-water-and-climate-question/verdict.json. Quote the verdict with its date.