Dholera vs King Abdullah Economic City: the port won alone

In 2005, Saudi Arabia announced a new city on the Red Sea coast, up the shoreline from Jeddah, and gave it the least modest name in the genre: King Abdullah Economic City. The scale was petro-capital at full stretch. A deepwater port, an industrial valley, resort districts, and a population of two million by 2020, delivered not by a ministry but by a company. Emaar The Economic City, the master developer, raised its money from the public: Saudipedia's company history records the IPO in July 2006, and the Saudi exchange's own profile shows the shares trading on Tadawul from October 2006 under the symbol 4220. A city, floated like a stock.

Two decades later, KAEC is the row that sits one place below Dholera in the Greenfield City Viability Index, the eighteen-city scoring exercise behind this site, and I think it is the single most instructive row on the interactive board for anyone holding a Dholera position. Not because KAEC failed the way internet lore says new cities fail. The opposite. KAEC built the hard thing, ran it to a world standard, and still did not become a city. If you want to stress-test the belief that Dholera's fab settles the argument by itself, this is the test rig.

The two rows, side by side

The index scores eight dimensions from 0 to 5 and weights them into a total. KAEC totals 2.55, ninth of eighteen. Dholera totals 3.10, eighth. The paper's verdict line on KAEC's anchor is blunt: the port was delivered to world-class standard, the industry and the city were not. Its status line is blunter still: restructuring under the sovereign fund, with roughly 10,000 people. The target it is scored against was 2,000,000 residents by 2020. Ten thousand against two million is half a percent, the thinnest attainment ratio of any case in the sample whose anchor delivery scores 3 or better.

Read the table twice and a strange thing happens. KAEC beats Dholera on the dimension everyone treats as the whole game, anchor delivery, and loses the comparison anyway. That inversion is this essay.

The port is real, and that is the point

Let me establish the anchor properly, because nothing here works if you imagine KAEC as a Potemkin project. King Abdullah Port is not a rendering. In the Container Port Performance Index for 2021, published in May 2022 by the World Bank and S&P Global Market Intelligence, it ranked first among 443 container ports worldwide. The same index is volatile year to year, and the 2022 edition placed it seventeenth, but a port does not visit the top of a global efficiency table by accident. Cargo moves, cranes work, shipping lines call, and they were doing so while most greenfield cities of KAEC's generation were still arguing with their own masterplans.

Industry came too, later and thinner than promised. Lucid opened AMP-2 inside KAEC's industrial valley on 27 September 2023, announced that day by the company and by Saudi Arabia's Public Investment Fund as the first car manufacturing facility ever commissioned in the kingdom. It started with semi-knocked-down assembly of about 5,000 vehicles a year from imported kits, against a stated full-build design capacity of 155,000 a year, which is its own small lesson in the distance between a plant opening and a plant at scale. That is a real factory with a real product, and it sits in exactly the category of headline that Dholera marketing would present as the arrival of destiny.

So KAEC is what a project looks like when the flagship infrastructure genuinely gets built and genuinely performs. Hold that thought against the resident count, which does not move: roughly 10,000 people in 2024, in a city drawn for 2,000,000.

Why a world-class port made almost no city

A port is a flow-through anchor. Value crosses its quays without needing to live beside them. A modern automated terminal employs a workforce that is small against the acreage it commands, and the professionals it does employ can sleep in Jeddah, a metropolitan labour market sitting down the coast. The index gives KAEC a 2 on proximate metro gravity, and in its case the nearby metro turned out to be a liability wearing an asset's clothes: close enough to staff the port, close enough to relieve the staff of any reason to move.

Then there is the sequencing. KAEC scores a 3 on demand anchor realism because the offer was an invitation, not an obligation: a freezone, incentives, a pitch. No instrument bound anyone to deliver two million people's worth of employment. The residential districts, meanwhile, were built and sold in the register petro-capital knows best, parcel by parcel, and homes marketed as investments attract owners rather than occupants. The result after twenty years is plain arithmetic: an operating port, a car plant assembling kits, resort quarters, and a population that would not fill a mid-sized Gujarati town.

The wider survey, which I have walked through in the eighteen-row essay, finds that a delivered anchor separates the cities that filled from the cities that stayed empty. KAEC is the amendment to that finding, and it deserves to be printed in the margin of every Dholera deck: delivery is necessary and it is not sufficient, because anchors differ in how many households they force into existence. Call that property settlement gravity. A container port has almost none. The question that should keep a Dholera investor awake is how much of it a semiconductor fab has.

A city with a share price

The second lesson sits in the funding column, where KAEC scores 2 on financing durability and Dholera scores 4, and this is the gap I would defend hardest. KAEC's developer is a listed company. Its money came from an IPO and then from selling the city it was building, which means the project's cash engine was investor sentiment about the project itself. That structure is self-reinforcing in fair weather and self-consuming in foul. When sales slowed, the damage went straight to the balance sheet.

The repair bill is public. On 8 September 2024, Arab News reported Emaar The Economic City's capital optimisation plan: SAR 8.7 billion in total, restructuring SAR 3.8 billion of bank debt, converting SAR 4 billion owed to the Public Investment Fund into equity, and opening a further convertible facility of up to SAR 1 billion from the same fund. Strip away the finance vocabulary and the sentence underneath is simple: the state stepped in to own the outcome, two decades after the market was invited to. Notice what did not need rescuing. The port company kept working. The sovereign fund rescued the city company.

Dholera's plumbing is different by design. The region exists under the Gujarat SIR Act of 2009. The builder is DICDL, a special purpose vehicle incorporated on 28 January 2016, owned 51 to 49 between the state and the centre. The activation area's trunk works were funded through centrally approved packages of Rs 2,784.83 crore with matching equity released, and the NICDC Delivery Monitoring Unit's report to DPIIT dated 30 June 2026 records those trunk works complete. Appropriated public money moves slower than an IPO and embarrasses governments when it stalls, which is precisely what makes it durable. The honest caveat: appropriations are only as permanent as the budgets that renew them, and Dholera's 4 is a grade on fifteen years of evidence, not a law of nature.

The rhyme with the fab, stated without comfort

Here is where I stop being fair to KAEC and start being useful to you. Dholera's case rests overwhelmingly on one anchor: the Tata Electronics and PSMC semiconductor fab, approved by Cabinet on 29 February 2024, financed under a fiscal support agreement signed on 5 March 2025 with a headline investment of Rs 91,000 crore, designed for up to 50,000 wafers a month at maturity. Civil work is reported past the halfway mark, first silicon is targeted for around December 2026, commercial production is reported for mid 2028, and no chip has been produced yet. I have argued in the fab essay that this single facility carries the whole thesis, and I stand by it.

KAEC's record is the uncomfortable rhyme. A project can deliver its flagship to a world standard and still house half of one percent of the people it drew on paper. Swap the port for the fab and the rhyme is close enough to hum: a capital-intensive, automation-heavy asset, nationally celebrated, operating behind a fence, surrounded by land priced as though the city were already there. The claim attached to Dholera's fab is more than 20,000 direct and indirect jobs, and the sanctioned plan gives the electronics sector a potential of 87,300. Those are employment numbers. KAEC is the demonstration that employment numbers and resident numbers are different species, a distinction I have measured for Dholera itself in the population essay, where its score on traction is a zero awaiting evidence.

And Dholera has its own Jeddah. The Ahmedabad expressway, reported inaugurated on 31 March 2026, puts the state's largest labour market 40 to 60 minutes away depending on whose report you read, and the spread in that figure tells you how young it is. I have called the road an economic rewiring in its own essay, and it is, but rewiring conducts in both directions. A fab engineer who can sleep in Ahmedabad and reach the cleanroom in an hour has no urgent reason to move their family to a district where the schools and hospitals are still mostly on the plan sheet. The road that makes the fab staffable is the same road that lets the city stay unpeopled for years. KAEC ran that exact experiment with a coastline and a metropolis, and it produced ten thousand residents.

What breaks the rhyme, and what would confirm it

The differences run in Dholera's favour on paper, and I want them stated as precisely as the warning. First, anchor type. A fab is not a quay. It runs continuous shifts, it accretes an ecosystem of suppliers, industrial gases, test and packaging units, and the sanctioned plan reserves 11,000 hectares of industrial land for exactly that accretion. Manufacturing clusters have historically made towns in a way container terminals have not since automation. Second, Dholera's demand anchor is contracted rather than invited: a cabinet approval, a fiscal support agreement, a named operator with its own balance sheet. That is why it scores 5 on anchor realism to KAEC's 3. Third, the trunk city is owned by a statutory SPV, not by a developer whose share price doubles as the city's mood ring.

Against all that, respect the scale of what KAEC's backers could summon and still not conjure. This was not a project short of money, land, state attention or engineering. It had all four in quantities Dholera will never see, and the residents did not come. Base rates should humble everyone here: in the sample of eighteen, exactly one industrial greenfield filled to megacity scale, and it had Hong Kong next door and a forty-year head start.

So the discriminating evidence at Dholera is not the fab's ribbon cutting. It is where the workforce sleeps. Tata is reported to be planning about 530 worker apartments on its own land: a small number, and exactly the right category of number, because occupied worker housing is the first statistic that converts jobs into residents. I watch it the way I watch everything here, as dated, checkable events: first silicon around December 2026, airport operations targeted for September or October 2026 against a slip history running back to 2010, school enrolments, and any census-grade count of people actually living inside the region. If this comparison moves you toward or away from a plot, the mechanics stay boring: check the scheme's GUJRERA registration where registration applies, and part with nothing until the title inside the SIR is clear and marketable.

KAEC settles one argument permanently. Nobody can claim any longer that building the anchor settles the city. The port won its half of the bet alone, and the quiet districts behind it are the other half, still waiting. Dholera enters the same wager with a better anchor, better plumbing, a worse land history and one rank more credibility, on a board where every city above it has already done the thing it has not. The fab can win and the city can still lose. Knowing that is not pessimism. It is the exact size of the remaining risk.

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