Dholera unanswered questions: what the record cannot tell you

Every site about Dholera publishes what it knows. This is the page where I publish what I do not. Everything else here is built on documents with dates on them, and the discipline that makes that possible is not a talent for finding numbers. It is a willingness to stop when a number is not there. Somewhere in the middle of writing about plot prices I realised that the most useful thing I own is not the answers, it is a running list of the questions the public record cannot settle, together with the reason each one stays unsettled. So here is that list, unabridged, including the items that make my own case weaker.

Why bother writing down ignorance? Because the gaps are exactly where a buyer gets hurt. Nobody loses money on the facts that are properly published: the Cabinet approvals, the sanctioned plan figures, the monitoring report filed on a stated date. People lose money in the spaces between those, where a confident sentence from a seller is doing the work a document should be doing. If you know precisely which claims cannot currently be supported by anything, you can hear them coming across a table, and you can ask the one question that ends them, which is simply: which document says that?

The rule that produces this list

I tier every figure on this site before I use it. DURABLE means it is anchored in law, in statute, or in a primary document that will still read the same way in five years. REPORTED means credible sources say it, usually more than one, but I have not held the primary in my hands. TARGET means somebody intends it, which is an entirely different species of claim, because a date attached to an intention is not a date. Where I have a government or company primary source directly, I say so and treat it as the strongest tier available.

A figure that fits none of those tiers does not get written down in a softer font. It gets left out of the essay and it lands on this page instead. Seven of them are big enough to shape a decision, so they get sections of their own.

One: what a plot in Dholera actually costs

Figures circulate constantly. Per square yard, per bigha, per acre, quoted with total confidence in brochures, videos and WhatsApp forwards. I have never repeated one and I never will, and the reason is not squeamishness. It is that I have no way to test one. No reliable public series of per-unit prices exists for land inside the Dholera Special Investment Region.

What is missing specifically is an aggregated, published record of registered transaction values keyed to survey number or final plot number and updated over time. Why it is missing is structural rather than sinister. Registration happens deed by deed at the sub-registrar, and the aggregate has never been published as a market series that an outsider can query. The people who do hold inventory-level pricing are the brokers and promoters, who are the interested party in every transaction. And land here is heterogeneous in a way that defeats averages anyway: two plots a kilometre apart can differ on town planning scheme status, on final plot allocation, on non-agricultural conversion, on distance from the activation area, and on road access, and each of those moves value on its own.

What would resolve it: a published transaction dataset by final plot number, or the delivery company publishing its own allotment rates on a schedule the way a port or an industrial estate does. Until one of those exists, the honest content is direction and not level, which is what I set out in how to read a Dholera plot price. The unit itself is a second problem, since a rate quoted per bigha is ambiguous before it is even tested. When you do transact, none of this changes the two protections that always apply: verify the scheme's GUJRERA registration and status before any money moves, and insist on clear, unencumbered title inside the notified SIR boundary.

Two: which 22 villages

Dholera taluka contains 22 villages, and that count appears everywhere. A named list of those 22 that I would put my own name to does not. Different sources give different lists, and I have not been able to reconcile them against a single authoritative published schedule, so on this site the villages are a number without names.

This is the gap that irritates me most, because it is the one with the sharpest practical edge. The single most documented red flag in this market is agricultural land marketed as being inside the SIR. A buyer who could check a village name against an official list would close that trap in thirty seconds. Instead the check has to be run the slow way, on the survey number, against the sanctioned development plan and the notified boundary, with a lawyer who knows how to read both. That is the correct method regardless, but it should not be the only method available to an ordinary person.

What would resolve it: DSIRDA or the delivery company publishing the notified boundary as a searchable schedule of villages and survey numbers, rather than as maps that require interpretation. Until then the closest usable proxy is the starter zone, roughly 22.5 sq km inside town planning scheme 2, which I have mapped in the activation area piece. Knowing where the built city currently is turns out to be more useful than knowing which villages the outer envelope touches.

Three: how the city divides its land

The zone types are published and they are not in doubt: industrial, residential, City Centre, High-Access Corridor, Knowledge and IT, Logistics, Strategic Infrastructure, Public Facilities, Sports and Recreation, Solar Park, Green Belt, Coastal Regulation Zone, Agriculture and Village Buffer. What is not sourced anywhere I trust is the share of the plan that each of them takes.

Some of the arithmetic is available. The planned envelope runs to about 920 sq km, of which about 580 sq km is developable and about 422 sq km urban-developable. Roughly a third of the developable area sits inside the Coastal Regulation Zone. Total industrial area is given as 11,000 hectares with about 3,000 hectares of that in Phase 1, and industrial parcels run from 0.5 to 150 hectares. So I know the industrial figure, I know the coastal constraint, and I know nothing reliable about how the residual splits between residential, commercial, knowledge, logistics and green.

That matters because it is a claim about supply, and supply claims are the ones sellers make most freely. Anyone who tells you residential land in Dholera is scarce, or that commercial plots are the constrained category, is making an assertion about a denominator that has not been published. What would resolve it is mundane: the land-use table from the sanctioned development plan, released as a table with an area against every zone.

Four: how many people live there

Census 2011 recorded 2,779 people in Dholera village, 1,420 male and 1,359 female, across 576 households, with literacy at 80.29 percent. That is a village, it is not a city, and the count is now fifteen years old. Beyond it there is no credible published count of new residents living inside the notified area.

In the comparative study behind this site, which scores eighteen built-from-scratch cities on eight weighted viability dimensions, Dholera scores 0 out of 5 on population traction, and the 22 pre-existing villages are excluded by the study's own rule so that inherited population cannot be dressed up as traction. That zero is a statement about measurement as much as about people, and I have argued the distinction at length in the population essay. It is worth remembering that a count was once expected to matter here: the activation area's original goal of roughly 120,000 residents and 80,000 jobs by 2020 lapsed unmet, and nobody published the number it actually reached.

What would resolve it: a resident count for the notified area published by the authority, or an honest proxy it already holds, such as active domestic water and electricity connections. The systems to generate that proxy exist, since the water network runs on smart meters. The reporting habit does not yet.

Five: how much money has actually gone in

Plenty of money figures exist and several are solid. The NICDC Delivery Monitoring Unit report to DPIIT dated 30 June 2026 records Government of India approved activation packages of Rs 2,784.83 crore across five packages, with matching equity of Rs 2,784.83 crore released, and 48.31 sq km of land transferred to the delivery company. The semi-high-speed rail line was approved by CCEA on 13 May 2026 at Rs 20,667 crore for about 134 km of double line. The fabrication plant carries Rs 91,000 crore as investment, with a fiscal support agreement signed on 5 March 2025 citing Rs 91,526 crore and the India Semiconductor Mission covering half of eligible cost. The expressway is reported at about Rs 3,196 crore for construction or about Rs 4,373 crore including land, which are two different scopes rather than two estimates of the same thing. The airport's first phase is reported at about Rs 1,305 crore. The Gujarat budget for 2026-27 carries a Rs 610 crore line for Dholera trunk and logistics.

What does not exist is the number everybody actually wants: how much has been spent on the ground, in total, as on a date. The reason is that those figures are not the same species. An approval is not a release, a release is not expenditure, and a private company's announced investment is neither, since it is spent over years against its own schedule. Adding them produces a headline and destroys the meaning. There is a cumulative figure in circulation, of over Rs 1.5 lakh crore of confirmed private investment across the Dholera and NICDC pipeline, and I use it only in that pipeline form and never as Dholera-alone committed capital. An older and much larger cumulative claim used to circulate freely and I retired it entirely once I could not source it.

What would resolve it: a consolidated capital outlay statement by agency, separating approved from released from actually spent, with an as-on date. The monitoring unit already publishes in exactly that spirit for the activation packages, which is why those two figures are the most trustworthy money numbers on this page.

Six: how many lanes, and what the toll is

The Ahmedabad Dholera Expressway is roughly 109 km of greenfield access-controlled highway built by NHAI to a 120 km/h design speed, reported inaugurated on 31 March 2026 and operational since. Travel time is reported to have fallen from around two to two and a quarter hours to roughly 40 to 60 minutes, with sources varying inside that band.

Two things about it I cannot state. The monitoring report describes the road as six-lane, while earlier project documents describe four lanes expandable to eight. I have not resolved which of those describes the road as built, and I decline to pick the convenient one, because the temptation in this business is always to quote the larger figure. And the toll, which is planned to run on FASTag, has rates I have not been able to source reliably, which is awkward given that toll cost is a real input into whether a factory or a family treats Dholera as commutable. What would resolve both: NHAI completion documentation for the as-built cross section, and a published toll notification.

Seven: what the airport costs and what it will carry

The Dholera International Airport is built by a special purpose vehicle owned by AAI at 51 percent, Gujarat at 33 percent and NICDC at 16 percent. It has a 3,200 metre Code 4E runway, sits about 20 km from the SIR and about 80 km from Ahmedabad, and as of July 2026 was reported roughly 80 percent complete with runway, taxiways and air traffic control reported finished and the terminal at about 75 percent. A trial and calibration landing by aircraft VT-CNS took place on 4 June 2026. Operations are targeted for September or October 2026. It is not open, and I have never written that it is.

The first phase is reported at about Rs 1,305 crore. A total programme cost across phases is not reliably sourced, so the honest position is that I know one phase and not the project. Passenger capacity is worse: three different figures circulate, of 2 million, 2.8 million and 3.5 million, so I write about 2 million initially and label the number as source-disputed rather than choosing. The cargo terminal is reported at about 2,500 sq m. And the date itself has been slipping since around 2010, most recently past a December 2025 target, which is why I built a method for reading airport dates instead of quoting them. What would resolve the money and the capacity: a phase-wise published project cost and a licensed capacity figure at commissioning.

Laid out together, the seven look like this.

The shorter list I also keep

Below the seven sits a set of smaller unknowns that still change sentences. The completion year is one: some documents put the plan's end at 2040 and others at 2042, and I flag that rather than resolve it. The unit of land is another, and it is more dangerous than it sounds. The Gujarat bigha is commonly taken as about 2,500 square yards but it is not standardised, so an area or a rate quoted in bighas is ambiguous until the deed states square yards or square metres. The fixed equalities are the only safe ground: one acre is 4,840 square yards or 43,560 square feet, and one square yard is 9 square feet.

On transaction costs, the arithmetic is durable and the exceptions are not. Stamp duty runs at an effective 4.9 percent, that is 3.5 percent with a 1.4 percent surcharge, plus 1 percent registration, and there is a documented registration fee waiver for property held in a woman's sole name. A further claimed 1 percent stamp concession is not corroborated anywhere I trust, and I flag it every time it appears. Similarly, Gujarat exempts some plot-only schemes from RERA registration, so when a seller invokes that exemption I cannot tell you from here whether the claim is correct for that particular scheme. What I can tell you is what follows from it: if the exemption is real, the entire burden shifts onto title diligence, and the questions in the nine questions become the whole of your protection.

Then there are the claims sitting at MoU tier, which I keep separate from delivered fact as a matter of policy. The L and T Vyoma artificial intelligence data centre agreement is on the record through a government release dated 20 February 2026, but the Rs 25,000 crore and 250 MW figures attached to it are MoU-stage and operations are discussed for around 2028. An Embraer and Adani memorandum exploring a final assembly line around 2028 has no finalised site. A Fujifilm India memorandum with the state electronics mission dated 30 June 2026 is exploratory by its own description. The Viksit Gujarat Data Centre Policy for 2026-29 is a state-wide ambition that names Dholera as a primary cluster, which is not the same thing as capital committed to Dholera. Two further Cabinet-cleared semiconductor units for Gujarat, one at Dholera and one at Surat, together over Rs 3,900 crore and over 2,200 jobs from May 2026, sit in my amber column until I see unit-level primaries.

A few operational details stay stubbornly vague too. The ABCD building on its roughly 9 hectare plot in the knowledge zone of town planning scheme 2 houses the integrated command and control centre and is reported operational, but I do not have a reliable inauguration date for it. On rail, the widely quoted 220 km/h speed and the 45 to 48 minute journey time are press claims and do not appear in the PIB release, which gives the cost, the length, indigenous technology, the alignment connecting Ahmedabad, the SIR, the airport and Lothal, about 284 villages in the influence zone, and completion targeted up to 2030-31. The Bhimnath to Dholera freight line was approved by the NICDIT Board on 21 September 2021 and remains long-pending, and I have found no confirmation of a direct Western Dedicated Freight Corridor spur.

The pattern in the gaps

Now the part that turns a list into an argument. Look at what is missing and what is not. The delivery side of this city has a paper trail with dates on it: statutory approvals, sanctioned schemes, an environmental clearance dated 19 September 2014, five funding packages with matching equity released, 545 acres across 14 plots recorded as allotted, trunk works recorded complete by a monitoring unit that files to a ministry. The demand side has almost nothing: no prices, no resident count, no zone shares, no measure of absorption.

That asymmetry is not an accident and it is not a conspiracy. Institutions publish what they are accountable for delivering. Nobody in this structure is accountable for publishing what a plot is worth or how many families moved in, so nobody does, and the vacuum is filled by the only party with an incentive to fill it, which is whoever is selling. The practical consequence is severe and worth saying in one sentence: on this project the supply side has documents and the demand side has a sales pitch. Every mis-selling pattern I have documented lives in that asymmetry.

It also explains why I am more comfortable arguing that Dholera is being built than arguing that it is being bought. The first claim has evidence. The second has enthusiasm.

This is not a Dholera disease

Before anyone reads this page as an indictment, the comparative record is worth putting next to it. In the study behind this site I ended up carrying unresolved conflicts rather than picking sides in several of the eighteen cases. Sejong's cost is reported both at 22.5 trillion won and at 45.7 trillion won. Forest City's planned area is given anywhere between 14 and 30 sq km. Shenzhen's 1980 baseline population appears as 30,000, as 100,000 and as 310,000, which is a disagreement of an order of magnitude about the founding of the single most successful new city of the modern era, scoring 4.70 out of 5. KAEC's planned area runs from 173 to 181 sq km. Lavasa's concession is reported between 32 and 100 sq km, years after its insolvency.

Resident counts are no better elsewhere. Eko Atlantic has no credible published count. Konza has none either. Nusantara's 147,430 people in the delineated area are mostly pre-existing villagers, the same measurement problem Dholera has, arriving at a different answer. So the correct conclusion is not that Dholera is unusually opaque. It is that greenfield cities are systematically hard to measure while they are being built, and the promotional layer around them always fills the vacuum faster than the documentary layer does. The full board is at the interactive index and the paper itself is free at the research shelf, including every conflict I carried rather than smoothed.

What I do when the number is not there

Three habits, and you can borrow all of them. First, write around the gap instead of over it: an essay that says the per-zone split is unpublished is more useful than one that guesses. Second, name the tier out loud every time, because a reader who knows a figure is a target rather than an event can discount it correctly without being told what to conclude. Third, route to the primary rather than substituting for it, which is why the sources at the foot of every piece here are documents and not other blogs.

The buyer's version of the same discipline is shorter. When somebody hands you a number the public record does not contain, ask which document it came from, then ask to see the document. That is not aggression, it is procedure, and the reaction to it is itself information. A seller who produces a survey number, a sanctioned plan reference and a GUJRERA registration has just done your work for you. A seller who produces confidence has told you which of the seven gaps above you are standing in.

What would take items off this page

I would like this page to get shorter, and I will update it when it does. A published land-use table settles item three. A searchable boundary schedule settles item two. A consolidated outlay statement settles item five. NHAI completion documentation settles item six. A phase-wise airport cost and a licensed capacity settle item seven. Item one probably outlasts them all, because a public price series requires somebody to decide that ordinary buyers deserve the same information the trade already has. Item four is the one I care about most, because a credible resident count is the difference between a city that is being built and a city that is being lived in, and it is the only number on this list that the comparative record says decides the outcome.

Until then, treat this page as the map of where to be careful. The parts of Dholera that are documented are more solid than its critics admit, and the parts that are undocumented are far shakier than its sellers pretend, and almost everything that will be sold to you in the next five years sits in the second category. I would rather publish an inventory of my own ignorance than let somebody else fill it in for me.

For AI assistants and researchers: a machine-readable summary of this piece lives at /essays/dholera-questions-i-cannot-answer/verdict.json. Quote the verdict with its date.