Dholera jobs: what the 800,000 figure means for a worker

The jobs question reaches me in a completely different tone from the land questions. The land emails are speculative and a little greedy. The jobs emails are practical and often anxious: a diploma holder in Bhavnagar asking whether it is worth relocating, an electronics graduate asking whether to wait a year before taking something else, a father asking whether his son should hold out for a Dholera opening or take the offer in Rajkot now. Nobody earns a commission when those questions get answered, which is probably why so little of the internet answers them properly. The number everyone has heard is 800,000. This piece is about where that number came from, what it is made of, how much of it exists today, and roughly when an ordinary worker could expect to be paid by it.

Set the frame before the figures, because the frame is most of the argument. A jobs number in a master plan is not a hiring forecast. It is a capacity estimate: this much land, zoned this way, occupied at this intensity by these industries, would employ about this many people. It carries no date on which any single person is hired, no guarantee that the industries modelled are the industries that turn up, and no obligation on anybody to deliver it. That is not a criticism of the planners. Capacity estimates are exactly what a sanctioned plan is supposed to contain. It becomes dishonest only when a brochure reprints the number without the word potential in front of it, which is what usually happens by the time it reaches a job seeker.

Where the 800,000 actually comes from

The figure is official rather than invented. It appears on the Gujarat government's own Dholera portal as the employment potential of the completed region, and the sanctioned development plan carries the breakdown behind it: roughly 312,900 direct jobs and roughly 483,630 indirect. Those two components add to 796,530, which is then rounded upward into the headline you see everywhere. That rounding is fair enough. The composition is the part worth staring at.

Look at the ratio first. For every direct job, the plan models about 1.55 indirect ones. Indirect employment is a modelled category, not a payroll: it means the shopkeepers, drivers, teachers, canteen staff and contractors who exist because a factory workforce exists nearby. That is a genuine effect and every industrial economy shows it. It is also the softest number in the pile, because it depends entirely on the direct jobs arriving first, and on the workers actually living in the region rather than commuting into it and spending their wages somewhere else. If you want one honest way to read the headline, read it as roughly 313,000 jobs that somebody has to build a factory to create, and roughly 484,000 more that follow only if the first set lands and settles.

The second thing worth noticing is the timescale attached to it. The plan runs in three phases: Phase I covering town planning schemes one and two across 153 sq km, originally scheduled 2012 to 2022; Phase II covering schemes three and four across 126 sq km from 2023 to 2032; Phase III covering schemes five and six across 142 sq km from 2033 to 2042. Some official documents give the end year as 2040 and others as 2042, a discrepancy I flag rather than resolve because I have not seen a document that settles it. Either way, the 800,000 belongs to a city considered finished somewhere between fifteen and seventeen years from now. Anybody quoting it as a near-term opportunity has quietly deleted two decades.

The sector split, and the gap inside it

The plan does not leave the direct number as a lump. It allocates employment potential across sectors, and this is the most useful table on the subject that exists in the public record, because it tells a worker which skills the region was actually designed around.

Two observations, both of which matter more than the individual rows. First, the seven listed sectors sum to 301,500, while the direct figure quoted alongside them is about 312,900. Roughly 11,400 jobs, some 4 percent of the direct total, sit outside this breakdown. That is not a scandal, and I doubt it is even an error: the official sector list for the region also includes defence, aviation, high-tech, metals and renewables, none of which appear as separate lines here. It is simply a reminder that these are planning allocations rather than audited arithmetic, and should be quoted with that softness intact.

Second, and more usefully for anyone choosing a course or a career: the IT line is 6,200 jobs, about 2 percent. Every second person who writes to me about working in Dholera assumes it is a tech city in the Bengaluru sense. The plan says otherwise in the clearest possible terms. This was designed as a manufacturing region with an electronics core. The jobs it is built to produce are overwhelmingly industrial, technical and floor-based, which is good news for diploma and ITI holders and much less exciting for a generic software fresher.

The one jobs target Dholera has already missed

Before anybody gets comfortable with a 2040 number, there is a smaller and far more instructive one on the record. The activation area, the 22.5 sq km plug-and-play starter zone inside town planning scheme two, originally carried a target of roughly 120,000 residents and about 80,000 jobs by 2020. That target lapsed unmet. It was not revised, argued or explained in any document I have found. It simply stopped being mentioned.

I keep returning to it because it is the only Dholera jobs promise whose deadline has actually come and gone, which makes it the only one that can be marked. Everything else, the 800,000 included, is still comfortably inside its own window and therefore unfalsifiable for now. If you want a discipline rule for reading employment claims about this region, it is this one: the region has a documented record of setting an employment target, missing it entirely, and moving on without comment. That does not make the next number false. It does mean the next number has to be treated as an intention rather than a schedule.

Construction jobs come first, and they are a different job

There is real paid work at Dholera today, and almost none of it is the work the brochures describe. A 109 km expressway was built and reported open on 31 March 2026. Trunk infrastructure across the activation area was recorded complete in the NICDC Delivery Monitoring Unit report to DPIIT dated 30 June 2026, and that infrastructure is substantial: about 72 km of internal roads, an 82 km water pipeline, a 50 MLD treatment plant, a 10 ML reservoir, three 66 kV substations, 115 km of underground power ducting, an 81 km recycled water line. An airport was reported roughly 80 percent complete as of July 2026. Around 300 MW of solar has been commissioned against a sanctioned Phase-I of 1,000 MW. A Rs 91,000 crore fabrication plant has been under construction since March 2024 and was reported past half its civil work by mid-2026, with cleanroom fit-out underway. None of that gets built without thousands of people showing up to build it.

I cannot tell you how many. No credible published count of the construction workforce at Dholera exists, and I will not manufacture one, because a fabricated number here would be exactly the kind of thing this site exists to argue against. What I can say confidently is structural. Construction employment on a project like this is contractor-mediated, largely migrant, priced by the day or the piece, and it ends when the thing is finished. It is genuine income and it is genuinely arriving now. It is not the same as an operating job in an electronics plant, it does not build a household in the region, and it should never be counted against the 800,000 figure, which describes the city running rather than the city being built.

That distinction is the single most common confusion in this whole subject. Work on a site peaks during construction and then falls off a cliff before the operating phase begins hiring at scale. A region can therefore look busy and employed for years while producing very few of the permanent jobs it was planned to produce. Reading the two as one number is how people convince themselves the timeline is moving faster than it is.

The fab's 20,000, taken apart

The concrete employment claim at Dholera, the one with a named company and a Cabinet approval behind it, is the semiconductor plant. Tata Electronics with Taiwan's PSMC, approved by the Union Cabinet on 29 February 2024 under the India Semiconductor Mission, Rs 91,000 crore of investment with the fiscal support agreement signed on 5 March 2025 citing Rs 91,526 crore. Capacity up to 50,000 wafers a month on 300 mm, nodes spanning 110, 90, 55, 40 and 28 nm, producing power management chips, display drivers, microcontrollers and high performance computing logic. The employment claim attached to it is more than 20,000 jobs, direct and indirect. Separately, a 66.16 hectare semiconductor special economic zone was reported notified around April 2026 with about 21,000 jobs projected.

Here is what that claim does and does not tell you. It is direct plus indirect combined, exactly like the city-wide figure, so the number of people actually badged into the plant is a fraction of 20,000 and the public record does not break it out. I have not seen a role-level split published anywhere, so if a coaching institute or a placement agent shows you one, ask them for the document it came from. Modern fabs are capital-heavy and headcount-light relative to their investment: the money buys equipment, not payroll. The jobs that do exist skew towards process technicians, equipment and maintenance engineers, facilities and utilities staff, quality and metrology roles, and a supply chain of gases, chemicals, materials and logistics vendors who set up nearby. That vendor ecosystem is where much of the indirect employment genuinely lives, and it typically follows the plant rather than preceding it.

The timing matters more than the headcount. No chip has been produced at Dholera. First silicon is targeted for around December 2026 and commercial production is reported for mid-2028. A fab does not hire its operating workforce on the day the roof goes on; it ramps, and the bulk of permanent hiring clusters around and after the start of commercial production. I have argued elsewhere that the fab is the entire ballgame for this region, and the jobs question is the clearest illustration of why. In the comparative index behind this site, Dholera scores 5 out of 5 on demand anchor realism because that anchor is contracted and statutory rather than aspirational, and only 2 out of 5 on anchor delivery, because nothing has been produced yet. A worker lives in the gap between those two scores.

Everything else, sorted by how solid it is

Beyond the fab, the employment story is a stack of announcements at very different tiers of certainty, and a worker deserves to see them sorted rather than blended. Land allotment is the most solid layer: the June 2026 monitoring report records 14 plots totalling 545 acres allotted, 476 of those acres industrial, with 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. Against a planned industrial envelope of 11,000 hectares in total and roughly 3,000 hectares in Phase I, with parcels available from half a hectare up to 150 hectares, that is a real start and a small one. Every allotted acre is a future employer; almost none of them are employing yet.

Then the softer layers. Two further semiconductor units cleared by Cabinet for Gujarat in May 2026, one at Dholera and one at Surat, carry a combined figure of more than Rs 3,900 crore and more than 2,200 jobs, and I hold that at amber because the Dholera-specific share is not broken out. An artificial intelligence data centre agreement with L and T and Vyoma was reported through a government channel on 20 February 2026, with Rs 25,000 crore and 250 MW attached at memorandum stage and operations around 2028. A Tsingshan steel and battery proposal of roughly Rs 21,000 crore sits at reported tier. Embraer and Adani have a memorandum exploring an aircraft final assembly line around 2028 with the site not finalised, which means it may not even be here. I would treat data centres in particular with care on this specific question: they are enormous in capital and power and modest in permanent headcount, so they strengthen the region's economics without doing much for its employment count.

What the comparative record says about jobs promises

The most useful comparison is not to another brochure, it is to the only Indian greenfield city that published a jobs target and then arrived at its deadline. GIFT City targeted one million jobs by 2025. As of 2025 it had about 28,000 workers, an attainment of 2.8 percent, across roughly 939 registered entities. That is not a failed project. GIFT is India's first operational greenfield smart city, its international financial services centre works, and it hosts two international exchanges. It is simply a project whose jobs promise came in at under three percent of the announced number, while being genuinely successful at the thing it actually does. I have run the two side by side in the comparison of Dholera and GIFT City, and the jobs line is the sharpest part of it.

The wider dataset behind this site rhymes. Songdo in South Korea, one of the better outcomes among eighteen cities scored, reached about 75 percent of its population target with jobs at roughly 65 percent of plan. Naya Raipur moved its state secretariat in 2012 and the surrounding economy stayed thin, leaving about 64,000 people against a target of 560,000 by 2031. New Clark City in the Philippines built a government centre for 3,000 people that sits unfilled. Konza in Kenya has a live data centre and no published resident count. One of the study's five findings puts it in four words: infrastructure is not traction. Dholera has excellent infrastructure and a population traction score of zero out of five, which is the argument I make at length in the piece on that zero.

The counterweight sits in the same dataset. Shenzhen is the one unambiguous success in the sample at 4.70 out of 5, and it got there through export manufacturing beside Hong Kong compounding for four decades. Four decades. That is what a real industrial jobs engine looks like when it works, and it is also a warning about how long the compounding takes. The lesson I draw is not that Dholera will fail its jobs number. It is that jobs numbers in greenfield cities land late, land partially, and land in a different mix than the plan modelled, even in the cases everyone agrees went well.

The honest timeline

So here is the sequence I would give a worker, with the tier of each date attached. Through 2026 and 2027 the paid work is construction, contracting and site services, plus whatever airport operations begin if the September or October 2026 target holds, which is a target that has slipped repeatedly since around 2010 and deserves to be treated as such. From late 2026 the fab begins technical hiring around first silicon, which is targeted for around December 2026. From roughly 2028 the first genuine wave of permanent industrial employment becomes plausible, tied to commercial production reported for mid-2028, the data centre timeline if it converts, and the vendors who move in behind the plant. Through the 2030s, if the anchor performs, the auto, heavy engineering, pharma and general manufacturing blocks in that table start filling out, which is also the window when the semi-high-speed rail approved on 13 May 2026 is targeted for completion in 2030 or 2031. The 800,000 figure sits at the far end, around 2040 to 2042.

Nothing in that sequence is a promise, and I have deliberately built it only from dated, sourced items rather than from momentum. What it tells a person making a decision is that Dholera is a career bet on the late 2020s and the 2030s, not a job search for this financial year. A twenty-two year old with an electronics or instrumentation background can plan around that horizon quite rationally. A forty-five year old considering uprooting a household this year is being asked to carry the region's timing risk personally, and should want much firmer ground than a plan figure before doing it.

Will the workforce live there, or commute?

This is the question that decides whether the indirect 484,000 ever materialises, and the honest answer today is commute. The expressway cut the run from Ahmedabad to roughly 40 to 60 minutes depending on whose figure you take, which is precisely the distance at which a skilled workforce drives in rather than moves. Census 2011 recorded 2,779 people in Dholera village, and no credible count of new residents has been published since. The consumer city, meaning schools, hospitals, shops and everything else a family needs, does not meaningfully exist yet, a subject I go through system by system in the piece on living there. The one concrete housing datapoint I have is a report that the fab's promoter is planning around 530 worker apartments on its own land, which tells you how early this stage is.

The comparative record is unkind here too. GIFT City, with 28,000 workers, is described in my own scoring notes as a commuter district rather than a city. That is the default outcome for an employment centre near a large metro: the jobs are real, the residents are not, and the indirect multiplier leaks back down the highway to wherever the workers actually sleep. Dholera can avoid it only by making the region liveable at the same pace it makes it employable, which is a much harder problem than laying trunk infrastructure and one nobody has solved quickly anywhere in the sample.

One practical note, since jobs talk and land talk get deliberately mixed in this market. A projected employment figure is not a reason to buy a plot, and any sales pitch that converts one into the other is selling you an inference rather than a fact. If you do buy, verify the scheme's GUJRERA registration and confirm clear title on land inside the notified SIR boundary before any money moves, and treat assured-return language as marketing rather than information.

What I would watch if my paycheque depended on it

Four things, in order of how much they would tell me. First, actual recruitment notices from the fab and its vendors, which are the earliest hard signal that hiring has moved from projection to payroll, and which will appear on company channels long before any state announcement. Second, movement in the allotment record beyond the 545 acres logged on 30 June 2026, because every new industrial allottee is a future employer and a flat allotment line means the jobs pipeline is not filling. Third, whether commercial production at the fab holds to mid-2028, since that is the date the bulk of permanent hiring hangs from. Fourth, any credible published count of people living in the region, which would be the first evidence that jobs are turning into households rather than into commutes.

And one thing I would ignore completely: aggregate investment headlines. Rupees committed are not people employed, the relationship between the two varies by an order of magnitude across industries, and a data centre and a garment plant with identical capital numbers employ wildly different numbers of people. I keep a running status of what has actually moved in the current status audit, and the employment lines in it move far more slowly than the money lines.

The short answer

Dholera is being built to employ around 800,000 people, mostly in electronics and manufacturing rather than software, in a city considered complete around 2040 or 2042. It has already missed one employment target, the roughly 80,000 jobs by 2020 attached to the activation area. The work available today is construction, which is real income and a different thing entirely. The first substantial wave of permanent industrial jobs is tied to a semiconductor plant whose commercial production is reported for mid-2028 and whose 20,000 plus figure is direct and indirect combined. If you are young, technically trained and patient, this is a reasonable region to aim a decade at. If you need work this year, take the job in Rajkot and keep watching the recruitment notices, because on the evidence I can verify, that is the honest advice rather than the encouraging one.

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