Dholera current status: the August 2026 audit, system by system

Somebody asked me last week for the current status of Dholera in one sentence, and I refused, politely, because the one-sentence version is precisely how this project gets misrepresented in both directions. Compress it optimistically and you get a finished smart city, which does not exist. Compress it cynically and you get an empty desert, which stopped being an accurate description somewhere around the time a Rs 91,000 crore semiconductor plant crossed the halfway mark of its civil construction. The honest status of Dholera does not fit in a sentence. It fits in an audit, so I have written one.

Everything below is the position as of 3 August 2026, taken system by system: governance, land, trunk infrastructure, the airport, the expressway, the railway, the fab, the solar park, the second-line anchors, the policy layer, and finally the population, which is the system almost everyone forgets to check. I find it helps to think of a city under construction as a patient with many organs. Asking how Dholera is doing produces feelings; asking how each organ is doing produces medicine. Every material claim here carries a date and an evidence tier, because a status without a date is a rumour and a status without a tier is a pitch. At the end I will hand over the part of this essay I actually care about: a section showing how to re-verify the whole audit yourself in about an hour, from primary documents, without taking my word for anything.

The standing disclosure, since status updates are the natural habitat of plot marketing: I sell no plots, broker nothing, and earn nothing from what you conclude. My only declared interest is having read the documents first.

The method: three tiers and a suspicion of verbs

Three tiers run through this audit, and I show them in capitals wherever a claim could reasonably be doubted. DURABLE means settled record: a statute, a Cabinet approval, a company incorporation, a census table, a signed fiscal agreement. Facts of this class do not move when a news cycle does. REPORTED means credible but not settled: progress percentages, completion claims, cost figures from official interviews and the business press, which I believe on balance but cannot pin to a primary ledger. TARGET means a date somebody intends to hit. Dholera's own history obliges me to say plainly that TARGET is the weakest tier of the three, because this project's targets have a documented habit of arriving late, and an audit that treats intentions as achievements is just marketing with footnotes.

Where a claim traces to a primary government document I say so explicitly, because that is the gold standard available here. The most important such document right now is the NICDC Delivery Monitoring Unit report to the Department for Promotion of Industry and Internal Trade, dated 30 June 2026, which functions as the closest thing this project has to a public ledger. You will meet it repeatedly below, and once more at the end, where I show you how to read it yourself.

One more instrument before we begin: watch the verbs. Complete, operational, approved, targeted, proposed, explored. Those six words describe six different quantities of reality, arranged in descending order, and half the confusion about Dholera's current status comes from claims being quietly promoted one verb up the ladder. Most of the value of what follows is nothing more sophisticated than putting every claim back on its correct rung.

The whole audit in one table

Here is the summary before the arguments. Each row gets its full section below, because a table cannot hold caveats, and in Dholera the caveats are where the truth lives.

Governance: the machinery exists, and it reports upward

Start with what Dholera actually is in law, because the marketing rarely bothers. Dholera SIR is a Special Investment Region constituted under Gujarat's Special Investment Region Act of 2009, which is DURABLE statutory fact rather than a branding exercise. The site sits about 100 km southwest of Ahmedabad, in Dholera taluka of Ahmedabad district, on the flat, low-lying Bhal plain along the Gulf of Khambhat. The planning envelope is large enough to mislead: roughly 920 sq km on paper, of which about 580 sq km is considered developable and about 422 sq km urban-developable, with roughly a third of the developable area inside the Coastal Regulation Zone. Those envelope figures are REPORTED from government planning documents, and I repeat them constantly because almost every exaggerated Dholera claim begins by quietly treating all 920 sq km as a city under construction. It is not. The action is concentrated in a small fraction of the map, and knowing where is most of the game.

Two bodies matter, and their division of labour explains most press releases. DSIRDA, the Dholera Special Investment Region Development Authority, is the planning authority. DICDL, Dholera Industrial City Development Limited, is the company that actually builds: incorporated on 28 January 2016, owned 51 percent by Gujarat through DSIRDA and 49 percent by the Centre through the NICDC Trust, all DURABLE. Above them, NICDC is the central nodal agency for the industrial corridor programme, and Dholera stands as the largest node of the Delhi-Mumbai Industrial Corridor, a superlative that appears in the official record. The ownership structure means two governments are financially wedded to the same balance sheet, which is worth something real: projects with one sovereign parent get orphaned when politics turns, and projects with two have a harder time dying quietly.

The status of the governance system itself is easy to state: functioning, and more importantly for an auditor, reporting. The Delivery Monitoring Unit report of 30 June 2026 exists precisely because this machinery files updates to a central department, and a project that publishes its own delivery ledger, with numbers a stranger can quote back at it, is a different animal from a project that communicates only through brochures. The planning skeleton beneath it: six town planning schemes, all sanctioned in draft, REPORTED, phased as 153 sq km for TP1 and TP2, then 126 sq km for TP3 and TP4 across 2023 to 2032, then 142 sq km for TP5 and TP6 across 2033 to 2042, summing to the urban-developable 422. The first phase carried a paper window of 2012 to 2022, which has already told you something about how paper windows behave here. I have walked the whole sanctioned plan end to end separately; for a status audit, the essential point is that the honest clock on this city is measured in decades, and the plan says so itself. One inconsistency I flag rather than resolve: the end year appears as 2040 in some documents and 2042 in others. When a project's own paperwork disagrees about its finish line by two years, the correct posture is a note and a shrug, not a confident citation.

Land: the ledger entries, and the litigation underneath them

Land is where Dholera's past is buried, sometimes literally, so I take the current entries first and the scar tissue second. From the 30 June 2026 DMU report, all primary: 48.31 sq km of land stands transferred to DICDL, the builder company. Fourteen plots totalling 545 acres have been allotted, 476 acres of that industrial, with Tata Chemicals named as the anchor industrial allottee. A further 1,043 acres of industrial land and 1,031 acres of other land are recorded ready for allotment. Environmental clearance for the development dates back to 19 September 2014. Read those figures the way an accountant would: small against 920 sq km, meaningful against the activation area, and above all entries in a register rather than lines in a press release. An allotment table with named companies is the kind of boring that I have learned to respect.

Now the history, because a status audit that hides the scars is a sales document. Land assembly here was litigated hard. The Gujarat High Court stayed SIR acquisition proceedings in 2015 after farmer petitions, and a 2017 review in the national business press found only around 290 of the more than 900 sq km then secured, in a year when the wider corridor programme was being written up as a story of delays and abandonments. That era is DURABLE history, not rumour, and it explains two present-day facts at once: why the project leans so heavily on the town planning mechanism, which pools land with landowners rather than simply acquiring it, and why title diligence in this market is not optional ceremony. Since buying is where land status meets your money, one standing line and no more: verify any marketed scheme on GUJRERA and insist on a clean, lawyer-checked title inside the SIR boundary before any money moves, and treat a seller's resistance to either check as your answer.

For orientation, the SIR spans 22 villages of Dholera taluka, though I have never seen a fully verified named list published and I decline to invent one. What matters for a buyer or an observer is not village poetry but planning geometry: inside a sanctioned TP scheme or outside it, inside the activation area or far from it, non-agricultural by rule or agricultural in fact. Everything else is brochure geography.

Trunk infrastructure: the finished part of the promise

If this audit has a headline, it lives in this section. The Activation Area is the roughly 22.5 sq km starter zone inside TP2, recorded as 22.54 in NICDC's own material, conceived as the plug-and-play patch where industry can arrive and connect rather than arrive and wait. The DMU report of 30 June 2026 records the trunk infrastructure works there as complete. The word in the filing is complete, not progressing and not substantially achieved, and it appears in a government body's report to another government body, which is the least theatrical genre of writing that exists. The same ledger records the money: Government of India approved activation packages worth Rs 2,784.83 crore across five packages, with matching equity of Rs 2,784.83 crore released. In a country where infrastructure announcements comfortably outnumber infrastructure, a sovereign ledger recording spend and completion is the strongest single card Dholera holds, and it is the reason my own tone about this project shifted from amused to attentive.

What did that money physically buy? The sanctioned plan and the project company's 2024 account, REPORTED tier with plan-document backing, inventory it precisely. About 72 km of internal roads, 18 to 70 metres wide, built with cycle tracks and a corridor reserved for future public transit. A 50 MLD water treatment plant, 100 MLD of potable water available, a 10 million litre reservoir, and 82 km of water pipeline running on smart meters with non-revenue water reported under 5 percent, a figure most Indian municipalities would frame and hang on a wall. A 10 MLD sewage treatment plant and a separate 20 MLD common effluent treatment plant for industry, with an 81 km recycled-water network so treated water goes back to work instead of into the creek. Stormwater engineered around a 6.5 km canal, which on this flat coastal plain is not landscaping but the difference between a monsoon and an insurance event. Power delivered through three 66 kV substations and about 115 km of underground ducting, meaning there is no overhead cable spaghetti to unbuild later. Solid waste handled by 25 tonnes per day of segregation capacity, 30 tonnes per day of bio-methanation, two incinerators of 25 tonnes per day each, and a 28 hectare landfill.

The command layer sits in the ABCD building, on a roughly 9 hectare plot in TP2's knowledge zone, LEED Gold by design, housing the Integrated Command and Control Centre that runs the city systems. It is reported operational, and I mark that REPORTED deliberately, because I have not found a primary record of its exact inauguration date and this audit does not round up. What I can say with confidence is that the utilities inventory above is the least glamorous and most persuasive thing about Dholera in 2026. India has plenty of townships that built towers and never finished the pipes. It has very few zones where the pipes sit finished and waiting for towers, and of those two mistakes, only the second one corrects itself.

Honesty requires the other entry too. The activation area carried an earlier promise of roughly 120,000 residents and 80,000 jobs by 2020, and that target lapsed unmet, a DURABLE lapse that no amount of present-tense completion erases. The infrastructure arrived late and the people have not arrived at all yet, and both halves of that sentence are true simultaneously, which is Dholera in miniature. For the buyer's version of this geography, where distance from the activation boundary becomes a decision variable, I keep a separate field manual on the activation area.

The airport: mostly built, repeatedly late, still shut

Now the system with the longest rap sheet. Dholera International Airport is being built by DIACL, owned 51 percent by the Airports Authority of India, 33 percent by Gujarat, and 16 percent by NICDC, a DURABLE structure. The runway is 3,200 metres and Code 4E, sized for wide-body aircraft, DURABLE specification. The site sits about 20 km from the SIR and about 80 km from Ahmedabad. All of that is settled. Everything else about this airport should be read with the file open, because the file is unflattering: opening targets have slipped repeatedly since roughly 2010, and the most recent casualty was December 2025, which was missed. That pattern is DURABLE, and it has earned the airport a permanent discount rate in my ledger.

Here is what changed in 2026, and it is genuinely material. On 4 June 2026, an Airports Authority aircraft with registration VT-CNS made the first trial and calibration landing on the new runway. Around that landing, the runway, taxiways, and air traffic control infrastructure were reported complete, the terminal about 75 percent done, and overall completion near 80 percent as of July 2026 reporting. Operations are now TARGETED for September or October 2026. A trial landing is not an opening, but it is not nothing either: you cannot fake a landed aircraft, and the kinds of delay that remain possible after one are different in nature from the kinds before it. My full argument lives in the essay on reading airport dates, and it compresses to this: the current window is the most credible target this airport has ever issued, and it still deserves the word targeted until a scheduled flight sells a ticket.

The numbers around the airport carry their own caveats, so I state them with labels attached. Phase-1 cost is reported around Rs 1,305 crore. Initial passenger capacity is genuinely disputed across sources, with 2 million, 2.8 million, and 3.5 million all in circulation; I use roughly 2 million initially and call the figure source-disputed, because pretending to precision I do not have is how status reports rot. A cargo terminal of about 2,500 square metres is reported. And one glamorous item stays firmly on the bottom rung of the verb ladder: Embraer and Adani have signed a memorandum exploring a final assembly line around 2028, with the site not finalized. The operative words are exploring and not finalized, and a memorandum is a document about intentions. File it as interesting and unbankable.

What would an operating airport actually change? Less passenger romance than logistics arithmetic: an airport twenty kilometres from an industrial zone changes the mathematics of high-value, low-weight cargo, of executive movement, and of the quiet signalling that multinational site selectors price into location decisions. But that paragraph belongs in the future tense until departure boards light up, and this is an audit, not a forecast.

The expressway: the claim you can test with a car

The Ahmedabad-Dholera Expressway is the one 2026 deliverable you can verify with a fuel tank. It is roughly 109 km of greenfield, access-controlled road built under NHAI with a design speed of 120 km per hour, DURABLE specification, and it was reported inaugurated by the Prime Minister on 31 March 2026 and has been operational since. The drive from Ahmedabad, formerly a grind of two hours or more, is now reported at somewhere between 40 minutes and an hour, with sources varying and traffic doing what traffic does. I hold the whole item at REPORTED only because my top tier requires a government ledger entry, but of all the REPORTED claims in this audit it is the least disputable, since thousands of strangers drive the road every week and none of them have complained that it is imaginary.

An auditor still files three caveats. Cost figures circulate in two scopes, about Rs 3,196 crore for construction and about Rs 4,373 crore including land, and mixing the scopes manufactures fake precision. The lane count is genuinely inconsistent in the record: the DMU report calls the expressway six-lane while earlier documents describe four lanes expandable to eight, so I flag the as-built count as unverified rather than pick a favourite, a small discrepancy that earns its keep as a reminder that even primary documents disagree with each other. Tolling is planned on FASTag, and I have found no reliably sourced rate card, so any confident toll figure you meet is somebody's guess wearing a uniform. The road's limits belong in the status too: it moves people to Dholera in under an hour, and it holds no opinion on whether they stay. That second problem belongs to the population section, and no pavement solves it.

The railway: a real approval wearing a distant date

On 13 May 2026, the Cabinet Committee on Economic Affairs approved the Ahmedabad-Dholera semi-high-speed rail project, and because this one comes with a Press Information Bureau release, the core facts are unusually clean: Rs 20,667 crore, roughly 134 km of double line from the Sarkhej side of Ahmedabad, indigenous technology, connecting the city, the SIR, the airport, and Lothal, with about 284 villages in the influence zone and completion targeted up to 2030-31. The approval is DURABLE and the date is TARGET, and keeping those tiers separate is the entire point of this section: an approval is a real, funded commitment to begin a process, and it is not a train. The press has decorated the project with a speed of 220 km per hour and journey times of 45 to 48 minutes; those figures do not appear in the PIB release, so I hold them at arm's length until an official document owns them.

Why rail matters differently from road, in one auditor's paragraph: an expressway moves cars, which mostly means it moves the car-owning class, while a high-frequency railway moves everybody, which is what turns an industrial zone's labour market liquid. The freight story is older and humbler: a Bhimnath-Dholera freight line was approved by the NICDIT Board on 21 September 2021 and remains long-pending, and I find no confirmed direct spur to the Western Dedicated Freight Corridor, so heavy goods ride on road for now. Track this system by tenders and earthwork rather than by press speed claims, because approvals age well only when contractors follow them.

The fab: the anchor tenant, audited without romance

The Tata Electronics fab is the most consequential entry in this audit, and I have argued at length that it is effectively the whole ballgame, so here I confine myself to status and tiers. The DURABLE layer: Union Cabinet approved the project on 29 February 2024 under the India Semiconductor Mission, with Tata Electronics building through Tata Semiconductor Manufacturing and PSMC as technology partner. The Fiscal Support Agreement signed on 5 March 2025 cites Rs 91,526 crore, the commonly used round figure is Rs 91,000 crore, and the Mission covers half the eligible cost. Design capacity is up to 50,000 wafers per month on 300 mm wafers, across nodes of 110, 90, 55, 40, and 28 nanometres, making power management chips, display drivers, microcontrollers, and high-performance computing logic. A jobs claim above 20,000, direct and indirect, rides on the approval documents.

The REPORTED layer: ground broke in March 2024, civil construction passed the halfway mark by mid-2026, and cleanroom fit-out is underway. Production reportedly begins at the 90 and 55 nanometre nodes before walking down to 28, which is not a retreat but the standard commercial logic of mature-node fabs. Around the plant, the ecosystem is filling in: a 66.16 hectare Tata Semiconductor SEZ notified around April 2026 with about 21,000 jobs projected, REPORTED; a partnership with ASML confirmed on Tata Electronics' own pages, which upgrades that particular claim from press story to primary source; and a reported plan for about 530 worker apartments on Tata's own land, which happens to be the first concrete housing signal anywhere in this audit. The TARGET layer: first silicon around December 2026, with commercial production REPORTED for mid-2028. And the sentence that keeps this section honest: no chip has been produced yet.

The file also holds a cautionary fossil, and I keep it visible on purpose. The Vedanta-Foxconn memorandum of 2022, a joint venture whose value was put around 19.5 billion US dollars, collapsed when Foxconn withdrew on 10 July 2023. It was never committed capital, and its short unhappy life is the permanent reason this site refuses to count memoranda as money. The Tata fab cleared every bar that arrangement never reached: Cabinet approval, a signed fiscal agreement, poured concrete. That difference in kind, not in enthusiasm, is why one project anchors this audit and the other haunts it.

Solar: three hundred megawatts you can meter, and a stuck seven hundred

The Dholera Ultra-Mega Solar Park is the quiet overachiever of the record, and also a tidy demonstration of every tier at once. The DURABLE part: Phase-I stands sanctioned at 1,000 MW under GPCL, and about 300 MW is commissioned and generating, built by Tata Power with 873,012 modules across 1,320 acres, an installation the company describes as India's largest on a single-axis tracker, a claim I pass along labelled as the builder's own. The commissioned figure carries the strongest backing available short of standing in the field, appearing through the Central Electricity Authority's records as of September 2025 and through the DMU reporting. The REPORTED and TARGET part: the remaining 700 MW is under development with a target of March 2027, and it is late for documented reasons, tariff disputes that went before the state electricity regulator, which I note without adjudicating. And the marketing part: 5,000 MW, the number that headlines brochures, is a long-range ambition attached to roughly 2030. It is not installed and not under construction at that scale, and it belongs in the same drawer as the million-resident city.

Two auditor's notes complete the section. The park was approved around April 2018, so the arc from approval to first generation here is your best available base rate for how Dholera schedules actually run: real delivery, several years behind the excitement. And the reason solar matters to the wider thesis is not decoration: an industrial city selling itself to fabs and data centres needs a firm power story, and 300 MW of metered generation next door argues that story better than any policy document can.

The second line: agreements, memoranda, and the discipline of not adding them up

Behind the fab stands a queue of names, and this is the section where an auditor earns the fee, because every entry sits on a different rung and the marketing loves to stack them into one big number. The strongest entry: the Vyoma AI data centre from Larsen and Toubro, where an agreement is confirmed through the government's own news service dated 20 February 2026, with operations aimed around 2028. The famous figures attached to it, Rs 25,000 crore and 250 MW, remain MoU-stage numbers, so the agreement is real while the arithmetic is provisional. Next: two further Cabinet-cleared semiconductor units for Gujarat from May 2026, one slated for Dholera and one for Surat, with combined investment above Rs 3,900 crore and more than 2,200 jobs, which I hold at REPORTED pending fuller documentation. Then the exploratory shelf: Fujifilm India signed a memorandum with the Gujarat State Electronics Mission on 30 June 2026 to explore a semiconductor materials base, with explore as the operative verb, and a Tsingshan steel and battery investment of around Rs 21,000 crore circulates at REPORTED tier.

On cumulative arithmetic I follow one rule and recommend it to everyone. The defensible headline is over Rs 1.5 lakh crore of confirmed private investment across the wider Dholera and NICDC pipeline, and the word pipeline is doing load-bearing work in that sentence, because the figure is not Dholera-only and should never be quoted as if it were. You will also meet India's national semiconductor tally, around a dozen sanctioned projects across the country, dressed up in Dholera marketing as if the whole programme lived here; that is the nation's scoreboard, not this city's, and borrowing it is the cheapest trick in the brochure. When someone hands you a single giant Dholera number, ask which entries are signed agreements, which are memoranda, and which are Cabinet approvals. The sorting takes a minute and usually settles the conversation.

The policy layer: tailwinds are real, and they are not capital

Two 2026 policy entries belong in this audit because they shape the probability field even though they build nothing by themselves. Gujarat's data centre policy for 2026-29 declares a state-wide ambition of Rs 6 lakh crore and 7.5 GW of capacity, with Dholera named as the primary cluster. Every word of that is a state TARGET rather than committed capital, and I read industrial policy the way I read a discount coupon: its value depends entirely on who shows up to redeem it. More modest and more concrete, the Gujarat Budget for 2026-27 carries a line of Rs 610 crore for Dholera trunk and logistics work, REPORTED from budget coverage. The pattern across both is consistent: the state keeps writing Dholera's name into documents that sit near money. That is genuinely better than the alternative, and it is still not the same thing as a customer with a purchase order.

The population: the system where nothing has happened yet

Now the section every seller skips, which is exactly why I give it a full heading. The 2011 census, still the last full count available, recorded 2,779 people in Dholera village: 1,420 male, 1,359 female, 576 households, literacy 80.29 percent. That is DURABLE, fifteen years old, and still the most honest population figure in circulation, because no new census has replaced it and no resident wave has arrived to embarrass it. The activation area's old promise of 120,000 residents by 2020 lapsed unmet. The full-city promotional targets, roughly a million residents and about 800,000 jobs at maturity, with a published split of about 312,900 direct and 483,630 indirect jobs, sit on the official state portal as plan figures, and a plan figure with a multi-decade runway is a direction, not a forecast.

Here is the honest frame for this system. Cities are the slowest of all infrastructure. Roads take years and fabs take a few more, but populations take decades, because a household moves only when jobs, schools, healthcare, and other households have already moved, a circular dependency that only patience unwinds. As of this audit the first genuine seeds exist: an anchor employer physically building, reported worker housing in the hundreds of apartments, an expressway that makes weekly commuting rational. What does not exist is the thing itself: streets with strangers on them, a market that stays open after dark, a school with a waiting list. I say this as measurement rather than mockery, because the resident count is the metric that will eventually separate Dholera from every failed greenfield comparison, and as of August 2026 it has not started moving. Anyone who tells you otherwise is quoting the future in the present tense.

What is not true, as of 3 August 2026

A status audit should file the negatives with the same care as the positives, so here are the claims circulating this season that the record does not support. The airport is not open; it has a trial landing, a completion estimate near 80 percent, and a target window. No chip has come out of the fab; first silicon is a December 2026 target and commercial production a mid-2028 report. The solar park is not 5,000 MW; it is about 300 MW live with 700 MW targeted and delayed. There are no reliable public per-plot prices in this market, which means every confident per-square-yard figure you meet is a sales instrument rather than a statistic, and this site does not print what it cannot source. Assured returns and guaranteed appreciation are marketing phrases with no documentary existence anywhere in the record. And the title of India's first smart city, in its precise operational sense, belongs to GIFT City, which has been running since 2015 with a financial-services regulator on site; Dholera's defensible superlatives are different and honestly sufficient: the flagship greenfield industrial smart city, the largest node of the corridor, the site of India's first major commercial fab, and the rare Indian city built trunk-first.

How to re-verify this audit yourself in about an hour

Everything above decays. Reports get superseded, targets land or slip, and a status essay dated August 2026 will quietly mislead somebody by next year if it is read as scripture instead of as a snapshot. So here is the audit's real deliverable: the primary reading list, with the minutes each item costs and what to extract from it. Run it twice a year and you will know Dholera's condition better than nearly everyone attempting to sell you a piece of it.

First, twenty minutes: the NICDC Delivery Monitoring Unit report, the PDF published on NICDC's own domain, currently dated 30 June 2026. This is the ledger. Extract four things: the trunk works status, the approved package amounts and released equity, the land transferred to the project company, and the allotment record with its acreage. When a newer edition appears, compare it against the old one; the deltas between editions are the closest thing Dholera has to quarterly results.

Second, ten minutes: the DSIRDA Sanctioned Development Plan on the official dholera.gujarat.gov.in portal. You are not reading all of it. You are confirming the skeleton: six TP schemes, the phase windows, the zoning framework, the utilities inventory. This is the document that turns brochure maps back into planning law, and ten minutes with it inoculates you against most map-based salesmanship.

Third, five minutes: the Press Information Bureau. Release 2260624 covers the rail approval with the cost and the 2030-31 horizon, and the PIB archive holds the fab's Cabinet approval record from February 2024. PIB prose is deliberately dry, which is exactly its value. When a claim you keep hearing is absent from the release, as with the rail speed figures, the absence is itself information.

Fourth, five minutes: the Tata Electronics newsroom for the fab and the ASML partnership, with the Tata group newsroom behind it. Corporate pages are self-interested by definition, but for confirming that a milestone or partnership is officially claimed at all, the primary corporate source beats every aggregator repeating it.

Fifth, five minutes: the official Dholera SIR portal's about pages, for the plan-level population and jobs targets, so you know which numbers are directions rather than deliveries. Sixth, five minutes: Tata Power's material on the 300 MW plant, for the solar state of play. Seventh, five minutes: a dated, sourced timeline of project events, which I maintain independently precisely because status confusion is a business model in this market; any dated wire will serve, provided every entry cites a primary source. Eighth, and only if someone is selling you something, five minutes on the GUJRERA portal to check the scheme's registration number, promoter, and status yourself. The total is roughly an hour, requires no logins and no brokers, and costs nothing.

The habit behind the list is small enough to carry anywhere: dates, verbs, tiers. Ask when a claim was made, which verb it uses, and what class of document carries it. Do that, and you will perform on any Dholera pitch the same operation this essay has performed on the whole project, in less time than a site visit takes to schedule.

The watchlist: what moves this audit next

Between now and the middle of 2027, six events will re-grade this page, and I commit in advance to how I will read each one. The airport's September or October window: an opening inside it converts the project's most mocked system into its most visible, while another slip hardens a sixteen-year pattern into a verdict. First silicon at the fab around December 2026: on time or reasonably close, the anchor thesis strengthens; drifting badly, and every downstream date, including the reported mid-2028 commercial start, inherits the drift. The solar park's remaining 700 MW against its March 2027 target. The railway's first visible tenders and earthwork, which are the honest tests of a young approval. The second line hardening, meaning any of the data centre figures, the Fujifilm exploration, or the May 2026 semiconductor units maturing from memorandum arithmetic into signed, sited fact. And the quietest number of all: any evidence of resident households, because apartment windows with lights behind them will eventually mean more than every megawatt in this audit combined. I grade these in public on the quarterly scorecard, where this audit goes to age transparently.

So the current status of Dholera, compressed into one honest paragraph now that the evidence is on the table: a funded industrial project with its starter-zone trunk infrastructure verifiably complete on a government ledger, one expressway you can drive today, one airport nearly built and not yet open, one railway approved and not yet begun, one enormous fab past halfway with no chip yet, 300 MW of live solar with another 700 stuck in disputes, a thickening file of agreements that are not yet capital, a supportive policy layer that is not yet customers, and a resident population that has not yet arrived. That is not a finished smart city, whatever the brochures imply, and it is not a mirage, whatever the cynics repeat. It is a construction site that has started keeping its promises, surrounded by a sales industry that keeps making new ones on its behalf. Audit accordingly, and date everything you are told.

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