Dholera delays: every slipped date, and what they predict

I keep a plain text file of every Dholera date I have ever been quoted, with the source next to it and the day I heard it. It now runs to several screens, and most of the entries are obsolete. That file is the reason I find the usual argument about this project unhelpful in both directions. The sales version treats every announced date as a delivery. The dismissive version treats every slip as proof that nothing is happening. Neither side does the boring work of separating the dates that were missed from the dates that were met, or of asking what kind of promise each one was before it broke.

So this is the slippage record, assembled only from documents with dates on them, followed by an argument about what the pattern predicts. One framing point before the list: a delay is not a scandal in infrastructure, it is the base rate. What matters is which category a slipped date belongs to, because some slips are engineering, some are procurement, some are litigation, and one entry here is not a slip at all but a cancellation. Lumping them together produces the two useless positions above. Separating them gives you something you can actually use.

The airport, and what sixteen years of almost looks like

Dholera International Airport is the best teacher on this list because its slip history is the longest and the most public. Original targets have been slipping since around 2010, which means the facility has been about to open for roughly sixteen years. The most recent published deadline to fall was December 2025. As of July 2026 the project was reported at roughly 80 percent complete, with operations now targeted for September or October 2026: a reported figure and a target, not a delivery. The runway is 3,200 m and Code 4E, the special purpose vehicle DIACL is held by the Airports Authority of India at 51 percent, Gujarat at 33 and NICDC at 16, and the site sits about 20 km from the SIR and about 80 km from Ahmedabad.

What changed in 2026 is the composition of the remaining work rather than the calendar. On 4 June 2026 an Airports Authority trial and calibration aircraft, registration VT-CNS, landed there, with runway, taxiways and air traffic control reported complete and the terminal at roughly 75 percent. That is a genuine milestone and it is not an opening, and the distance between those two sentences is where most Dholera marketing lives. Phase-1 cost is reported at about Rs 1,305 crore, and even the passenger capacity claim is unsettled: figures of 2 million, 2.8 million and 3.5 million circulate, so the honest phrasing is about 2 million initially, source-disputed. I have written the full method for reading these announcements in the piece on how to read airport dates, and the short version is that a date attached to a terminal fit-out is worth more than a date attached to a ribbon.

The target nobody quotes any more

The quietest miss on the record is the one I never see in a brochure. The activation area, the 22.5 sq km starter zone inside town planning scheme 2, carried an original goal of roughly 120,000 residents and 80,000 jobs by 2020. That target lapsed unmet, and it did not lapse loudly. It simply stopped being repeated, replaced in the sales conversation by the promotional full-city figure of about one million people at maturity. Anyone quoting me the maturity number has to explain the 2020 number first, because it is the only population promise in this project's history that has already come due.

The maturity date itself is unsettled in the source documents, which appear to say 2040 in some places and 2042 in others. I flag that rather than resolve it, because picking the convenient year is exactly the habit this post exists to argue against. Meanwhile the measured population fact is Census 2011, which recorded 2,779 people in Dholera village, and there is still no credible published count of new residents in the SIR. A city can be behind schedule on infrastructure and still be building. A city with no resident count is not behind schedule on people, it is not yet measuring them, which is a different and more uncomfortable position.

The first decade, and why the plan window closed before the work did

The sanctioned plan splits the SIR into three phases: Phase I as town planning schemes 1 and 2, 153 sq km, planned for 2012 to 2022; Phase II as schemes 3 and 4, 126 sq km, 2023 to 2032; Phase III as schemes 5 and 6, 142 sq km, 2033 to 2042. Compare that first window with the delivery record and the slippage is plain. Trunk infrastructure works in the activation area were recorded complete in the NICDC Delivery Monitoring Unit report to DPIIT dated 30 June 2026, four years after the Phase I window closed, and the activation area is 22.5 sq km of the 153.

The reasons for the lost decade are documented rather than mysterious. In 2015 the Gujarat High Court stayed acquisition in the SIR after farmer petitions. In 2017 Business Standard reported that only around 290 of the 900 plus square kilometres had then been secured, and the wider corridor programme was described in the business press the same year as a tale of abandonments and delays. Land assembly through litigation is slow by design, and a planning authority that keeps working through a court stay is doing the unglamorous thing correctly even while the calendar bleeds.

What the same monitoring report records on the other side of the ledger is also dated and specific: 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, 48.31 sq km of land transferred to the delivery company DICDL, 14 plots covering 545 acres allotted of which 476 acres are industrial, Tata Chemicals named as the anchor industrial allottee, and a further 1,043 acres of industrial land plus 1,031 acres of other land ready for allotment. Environmental clearance was granted on 19 September 2014 and has held since. That is what a late project that is still moving looks like, and I set out how to use that starter map in the activation area essay.

The fab clock, which has not slipped yet

The most important clock on the site is also the youngest, which is why it has no slip history to report and why that absence proves less than people want it to. The Tata Electronics and PSMC fabrication plant was approved by the Union Cabinet on 29 February 2024 under the India Semiconductor Mission, with a fiscal support agreement signed on 5 March 2025 citing Rs 91,526 crore against the widely used Rs 91,000 crore headline, the Mission covering half the eligible cost. Groundbreaking followed in March 2024. By mid-2026 civil work was reported past 50 percent with cleanroom fit-out underway. First silicon is targeted around December 2026 and commercial production is reported for mid-2028. The dedicated 66 ha semiconductor special economic zone was notified around April 2026.

Two things in that paragraph get collapsed constantly. The first is the roughly eighteen month gap between first silicon and commercial production, which is not padding but the qualification and yield ramp that decides whether a plant sells anything. The second is that no chip has been produced at Dholera. The precedent that should keep everyone honest is the Vedanta and Foxconn venture announced for Gujarat in 2022, which was never committed capital and ended when Foxconn withdrew on 10 July 2023. That was a cancellation, not a delay, and the difference is the whole argument of this post. I have made the case that the fab is the ballgame precisely because its clock is the one that reprices everything else if it holds or breaks.

Solar, where the delay has a documented reason

The Dholera Ultra-Mega Solar Park was approved around April 2018 with a sanctioned Phase-I of 1,000 MW under GPCL. Roughly 300 MW is commissioned and generating, built by Tata Power with 873,012 modules across 1,320 acres, a figure that holds up through the Central Electricity Authority record of September 2025 and the monitoring report. The remaining 700 MW is under development with a target of March 2027, and the delay is attributed to tariff and regulatory disputes before the state commission rather than to construction. The 5,000 MW number that appears in promotional material is a long-range ambition around 2030 and is not installed capacity.

I find this the most instructive slip on the list, because a tariff dispute is a slower failure mode than a construction problem and a far more fixable one. Nobody is arguing about whether the panels can be built. They are arguing about the price at which the power is bought, which is a solvable commercial question that will get solved when the discount someone wants meets the return someone else needs. Eight years to deliver 300 of a sanctioned 1,000 MW is nonetheless eight years, and the arithmetic belongs in your model rather than in a footnote. The full accounting is in the solar park numbers essay.

What actually landed, and what that tells you

The other half of the ledger deserves the same precision. The Ahmedabad to Dholera expressway, about 109 km of greenfield access-controlled highway built by NHAI to a 120 km/h design speed, was reported inaugurated on 31 March 2026 and operational, cutting the drive from roughly two hours to somewhere between 40 and 60 minutes depending on which account you read. Even here the record has a soft edge worth naming: the monitoring report calls it six lanes while earlier documents describe four lanes expandable to eight, so I treat the as-built lane count as unverified. Cost is reported at about Rs 3,196 crore for construction or about Rs 4,373 crore including land, which are conflicting scopes rather than conflicting facts.

The semi-high-speed rail line is the newest entry and the cleanest example of a date that has not yet had a chance to slip. The Cabinet Committee on Economic Affairs approved it on 13 May 2026, recorded in PIB release 2260624 at Rs 20,667 crore for about 134 km of double line connecting Ahmedabad, the SIR, the airport and Lothal, with completion targeted up to 2030 or 2031. The 220 km/h speed and 45 to 48 minute journey figures that dominate the coverage are press claims and do not appear in that release, so I keep them at arm's length. The older Bhimnath to Dholera freight line, approved by the NICDIT Board on 21 September 2021, remains long pending, which is the quieter and more typical rail story here.

Laid out by date, the record reads like this.

Three kinds of slippage, and why lumping them together is lazy

Read that table twice and the entries sort themselves into three piles. The first pile is promotional targets that were never underwritten by anything: the 2020 activation population, the one million maturity figure, the 5,000 MW solar ambition. These are aspirations with a year attached, and treating their non-arrival as a broken promise gives the promiser too much credit, because a promise requires a promisor who was bound.

The second pile is funded delivery that arrived or is arriving late: the airport, the trunk works, the remaining solar, the phase windows themselves. Everything in this pile has money, a contractor and a paper trail, and everything in it has slipped by years rather than by decades. The third pile has exactly one entry, the Vedanta and Foxconn collapse, and it is the category that actually hurts, because a cancelled anchor cannot be waited out. When somebody tells you Dholera is behind schedule, the only useful reply is to ask which pile they mean.

What the pattern predicts

My working rule after assembling all of this is uncomfortably simple. Where there is a signed contract, a released budget tranche and an institution filing dated reports, the thing arrives, late, by a margin measured in years. Where there is an announcement and a rendering, the date is ordering information rather than scheduling information: it tells you what the state intends to do next, not when you can plan around it. Every entry in the second pile above had the first set of properties. Every entry in the first pile had the second.

That rule has company in the comparative record. In the study behind this site, an original index scoring eighteen built-from-scratch cities on eight weighted dimensions, Dholera scores 5 on demand anchor realism and only 2 on anchor delivery, for a weighted total of 3.10 and eighth place of eighteen. The gap between those two scores is this entire post expressed as a number. The sample's first finding is that the top seven cases are every case with a substantially delivered anchor and the bottom five are every case whose anchor failed or was never real. Saudi Arabia's KAEC targeted 2 million residents by 2020 and had about 10,000 in 2024. NEOM's Line has been cut in stages from 9 million toward roughly 100,000 by 2030. Sejong, by contrast, hit 78.9 percent of its target because 44 agencies and about 14,000 civil servants were moved by statute rather than by invitation. Missed dates are the norm in this category. What separates the survivors is whether the anchor was legally obliged to show up. The full board is at the Greenfield Index and the paper is free on the research shelf.

Applied to Dholera, the prediction is narrow and I will stand behind it. The airport opens, later than the current target more often than not, and the September or October 2026 window is the most credible one it has ever had because the remaining work is a terminal rather than a runway. The fab produces silicon within a year or so of its stated target, because Rs 91,000 crore of committed capital with a signed fiscal support agreement does not idle. The 700 MW of solar arrives once the tariff argument ends. And the population target is the one I would not bet on at all, because nothing in the machinery obliges a single household to move, which is exactly why the index scores Dholera zero on population traction while scoring it well on almost everything else.

What this changes if you are buying

Slippage is not an abstraction for a buyer, it is carry. Land here yields nothing while you hold it, transaction costs are real in both directions at an effective 4.9 percent stamp duty plus 1 percent registration, and every year a date moves is another year of holding an asset that pays you nothing. That is the actual mechanism by which delay hurts an individual, and it is why I think horizon discipline matters more here than entry timing. If your plan cannot survive the airport opening in 2027 instead of 2026, or commercial chip production in 2029 instead of 2028, then the plan was never calibrated to this project's demonstrated behaviour. The full ledger of what else can go wrong sits in the risks nobody lists.

The protections that survive any timeline are the boring ones. Verify the scheme's GUJRERA registration and status before any money moves, insist on clear title inside the notified SIR boundary with the chain of deeds and a thirty year encumbrance certificate behind it, and treat any assured return or guaranteed appreciation language as marketing rather than fact. No reliable public per-unit price series exists for Dholera land, so anyone quoting one is quoting themselves. None of that changes because a date moved. All of it becomes more valuable the longer the wait runs.

The dates I am holding this against

Predictions are only worth reading if they can be marked. So here are the six I will check, in the order they fall due: airport operations in September or October 2026, first silicon at the fab around December 2026, the remaining 700 MW of solar in March 2027, fab commercial production in mid-2028, the semi-high-speed rail line in 2030 or 2031, and the Phase II town planning window that runs to 2032. If four of those six land within a year of their stated date, the second pile is behaving exactly as the record says it should and I will say so. If two or fewer land, my rule needs rewriting rather than defending.

There is a seventh item with no date on it at all, and it is the one I watch hardest: a published, credible count of people living in the SIR. Nobody has committed to producing that number, which means it is the only line on this whole ledger that cannot slip, because it was never scheduled. A city is late until the day it is inhabited. Everything else on this page is engineering, and engineering, eventually, arrives.

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