The Future Wasn't Impossible. It Was Out of Order.

Essay · August 2026 · 21 sources, each marked doc or interp · how it was made

NEOM's collapse is being read as proof that cities of tomorrow can't be built. China has been building them for forty years — by refusing to build them the way Saudi Arabia did.

An essay submitted for the opinion pages. Every factual claim carries an endnote; the appendix separates documented fact from the author's interpretation.

In 1993, Milton Friedman stood in Shanghai's Pudong district — then a muddy expanse of state-directed construction across the river from the Bund — and pronounced it a monument in the making "for a dead pharaoh," a Potemkin skyline that markets would never fill.[1] It was a reasonable verdict. It was also wrong. Pudong today is the financial core of a city of registered exchanges, foreign banks and a skyline so familiar it functions as visual shorthand for the Chinese century.

Thirty years later, the same verdict is being rendered on NEOM — this time with the weight of evidence behind it. The Journal's own reporting has documented the wreckage in forensic detail: a linear city announced at 170 kilometers reduced to 2.4 kilometers of foundations[2]; more than $50 billion committed against an internal audit — prepared with McKinsey & Co.'s assistance — projecting as much as $8.8 trillion to finish by 2080, roughly twenty-five times the kingdom's annual budget[3]; auditors flagging "evidence of deliberate manipulation" in the financial models[4]; a resort island that opened three years late at triple its budget[5]; an executive instructing consultants that "we must not proactively mention cost at all"[6]; a chief executive gone by November 2024[7]; construction paused, then formally pared back.[8]

From this record the emerging conclusion — the one the documentary treatment lands on — is that the vision itself was impossible. The desert utopia failed because desert utopias fail.

That conclusion is comfortable, and it explains too little. Across the South China Sea sit a dozen cities every bit as hubristic as NEOM — cities that were mocked on these terms, in this newspaper's pages and on American television, a decade before NEOM broke ground. They are full now. If "impossible vision" is the diagnosis, someone must explain why the same disease killed the patient in Tabuk and spared it in Zhengzhou.

The difference is not money, not authoritarian will, and not vision. Both states command sovereign resources; both can flatten mountains by decree; both produced renderings that made architects blush. The difference is sequence. China builds cities the way engineers build bridges — demand first, structure second, spectacle last, corrections quiet. Saudi Arabia built NEOM the way studios build franchises — brand first, then a desperate search for inhabitants to cast in it. The future was never impossible. It was out of order.

The ghost cities that weren't

Consider the two most famous "failures" of Chinese city-building. In March 2013, 60 Minutes toured Zhengdong New District in Zhengzhou and broadcast empty towers and vacant subdivisions to America as the definitive Chinese ghost city.[9] Ordos Kangbashi, on the Inner Mongolian steppe, became the genre's photographic cliché — grand plazas, no people, roughly 30,000 residents rattling around a district built for hundreds of thousands.[10]

Zhengdong's government now counts 1.3 million residents.[11] The turn began within months of the 60 Minutes segment, and it had nothing to do with the skyline: Zhengzhou's first metro line opened in December 2013 and stitched the new district to the old city, while Foxconn's plant — which assembles roughly half the world's iPhones — anchored the region's payrolls, and the provincial government moved its own offices in.[12] Population in the district grew 27.5% in a single year at the peak; property values rose roughly tenfold over the decade.[11] Kangbashi, the harder case — genuinely remote, genuinely speculative — still filled to about 130,000 official residents by 2024, after the city government did the unglamorous work of relocating its best schools and administrative functions there and letting a decade pass.[10][13]

The pattern generalizes, and it has a name in the scholarship of Wade Shepard and others who bothered to revisit the ruins: China's ghost cities were not failures of vision but errors of timing, and timing errors are curable when demand exists to cure them.[13] The cure was never the architecture. It was a subway line, a factory, a relocated bureaucracy, a good school — the boring gravitational mass that gives people a reason to be somewhere.

The wave and the surfboard

Behind every cured ghost city stood the largest migration in human history. In 1980, 19.4% of China's population was urban. By 2024 the figure was about 67% — on the order of 800 million people moving to cities inside four decades.[14] Every new district, however speculative, was a surfboard on that wave. Build too early and you look foolish for five years; the wave arrives anyway.

Shenzhen, the founding myth of the genre, is remembered as a fishing village turned metropolis. The truth is only slightly less dramatic — a border town whose urban core held fewer than 60,000 people in 1979 — and the mechanism is instructive: Beijing did not decree a finished city. It declared a special economic zone, a legal experiment, and let firms, migrants and iteration do the designing. Today Shenzhen holds nearly 18 million people and produces roughly $530 billion in annual output — more than the GDP of Norway — without any founding master plan promising a mirror-clad utopia.[15]

Saudi Arabia has no wave. When NEOM was announced in 2017, the kingdom was already more than 80% urban — 85% today — one of the most urbanized societies on earth, with a native population of about 35 million whose city-dwellers were being asked not to become urban but to relocate from Riyadh and Jeddah, cities the same government was simultaneously spending billions to improve.[16] NEOM was a surfboard on a flat sea. That is not a moral failing, but it has an unforgiving arithmetic consequence: every resident had to be conjured — by brand, by amenity, by press release — rather than received. China has never successfully conjured residents either. Nobody has.

The control group

Here the argument usually collapses into essentialism — China succeeds because it is China. The record is more interesting: the variable is demand, not nationality, and China itself supplies the controls.

When a Chinese developer built a NEOM — a $100 billion instant city for 700,000 on artificial islands in Malaysia, complete with renderings of vertical forests — it produced Forest City: about 9,000 residents by late 2025, occupancy estimated at 15–30%, widely ranked among the great real-estate failures of the century.[17] Same builders, same capital, same confidence. No wave.

And when Beijing itself abandoned its own playbook — when it decreed Xiong'an, President Xi's flagship "city of the future," a top-down showpiece a hundred miles from any organic demand — it got NEOM-like symptoms at NEOM-like scale: roughly $119 billion invested by early 2025, tenanting achieved mainly by ordering state-owned enterprises to relocate their headquarters, and independent reporting in 2026 branding it a ghost city, with occupancy estimates circulating below 30%.[18] State media counts 1.4 million residents, but the three rural counties absorbed into the zone held roughly a million people before the first cornerstone was laid.[19] The exception proves the rule with uncomfortable precision: China's most Saudi-style project is China's least successful one. The lesson was never "China can and Saudi Arabia can't." It is that demand-following city-building works and demand-inventing city-building doesn't — for anyone.

The experiment Riyadh already ran

Which brings us to the detail the "impossible vision" verdict must ignore to survive: Saudi Arabia had already run the NEOM experiment, at merciful scale, and read the results before writing the check.

King Abdullah Economic City, launched in 2005 on the Red Sea coast with a master plan for two million residents and a million jobs, had attracted roughly ten thousand residents by 2024 — one-half of one percent of plan — despite a functioning port, a listed developer and two decades of runway.[20] KAEC failed exactly the way NEOM is failing: a city designed as a product, marketed to inhabitants who had no economic reason to come. A state that metabolized evidence would have treated KAEC as China treated its early zone experiments — as cheap tuition. Instead the response was to scale the failed design by two orders of magnitude and armor it against criticism by making it the personal brand of the crown prince.

That armoring is the true pathology, and it is institutional, not visionary. Zhengdong could be quietly rescued by a subway because no paramount leader's face was welded to its skyline; Chinese planning treats new districts as a portfolio — hundreds of bets, many failures, all correctable in silence, with local officials promoted on delivered growth rather than delivered renderings. NEOM was a single bet that could not be marked to market, reported through consultants whose revenue — McKinsey's alone exceeded $130 million a year — depended on the mark never being taken, inside a court where, as the audit found, managers inflated hotel-rate projections rather than carry bad news upward.[3][4][21] The fraud, waste and abuse the Journal documented are real. But they are the symptom of a structure with no error-correction channel — the same structure that made KAEC's lesson unhearable. Corruption did not kill NEOM. NEOM's design made honesty unaffordable, and dishonesty did the rest.

What a possible NEOM looked like

The tragedy is that a demand-following NEOM was available, on the same coastline, under the same name. Thirteen percent of global trade passes through the Suez Canal within sight of NEOM's shore. The kingdom holds structural advantages in energy — including the cheap solar power that data centers and green-hydrogen plants actually relocate for — and a genuine logistics gap between Europe, Africa and Asia. A Chinese-sequenced NEOM starts ugly: a port and free zone at Oxagon's site, an energy-industrial anchor, worker housing that looks like Shekou in 1982 rather than Blade Runner, a rail link, a school worth moving for. It lets the first 50,000 residents arrive for jobs, not press junkets. It banks two decades of boring compounding — and then, on top of proven demand, it builds whatever mirrored spectacle the sovereign's heart desires, as Shanghai eventually built its own towers on Pudong's proven base. The reported pivot of NEOM's remnant toward AI data centers suggests someone in Tabuk has finally located the demand curve[8] — a decade late, a half-trillion riyals short, but directionally the first Chinese sentence in the project's vocabulary.

The Line was not impossible. The Line was scheduled first when it should have been scheduled last. Friedman's pharaoh line was wrong about Pudong because Pudong was a bet on a wave that existed; it lands on NEOM because NEOM was a bet on a wave that didn't. Cities are not products that ship. They are verbs that conjugate — slowly, in the boring tenses first. The kingdoms that learn this build Shenzhens. The ones that don't build very beautiful, very empty proofs that the future, like the past, cannot be purchased in a single transaction.

Provenance appendix

Lane discipline, borrowed from the systems this essay was verified on: entries marked doc rest on published documents and reporting; interp flags the author's inference. Every bracketed number in the text resolves here.

1. doc Friedman's 1993 Pudong remark — a monument "for a dead pharaoh" — and its afterlife: The American Scholar, "Milton Friedman's Misadventures in China"; The National Interest, "The Visionary Milton Friedman and China".
2. doc The Line reduced from 170 km announced to a 2.4 km initial build: New Civil Engineer, Mar. 2025; Bisnow, "Saudi Megaproject Neom Could Cost $8.8T".
3. doc Internal audit (McKinsey-assisted) projecting up to $8.8T by 2080, ~25× the Saudi annual budget; >$50B committed: Engineering & Technology (IET), Mar. 10, 2025, reporting on WSJ's findings; Bisnow.
4. doc Auditors' phrase "evidence of deliberate manipulation"; inflated hotel-rate projections (e.g., rooms repriced $489→$1,866): E&T summary of WSJ audit reporting.
5. doc Sindalah island: opened Oct. 27, 2024, three years late, ~$4B against ~$1.3B original budget: Wikipedia: Sindalah (aggregating WSJ and trade reporting); Fast Company.
6. doc Antoni Vives email, "we must not proactively mention cost at all," per WSJ-obtained correspondence: Futurism, summarizing WSJ.
7. doc CEO Nadhmi al-Nasr's departure, Nov. 2024: Deadline; Wikipedia: Nadhmi Al-Nasr.
8. doc Construction pause reported late 2025; formal scale-back and reported pivot toward AI/data-center uses, Jan. 2026: Euronews, Jan. 26, 2026; Ecoticias (English), on the AI-infrastructure pivot.
9. doc 60 Minutes "ghost city" segment on Zhengdong, Mar. 2013: Vagabond Journey field reporting; Newgeography, "Zhengzhou Ghost City Alive!".
10. doc Kangbashi: ~30,000 residents circa 2009; ~130,000 official by 2024: Wikipedia: Kangbashi District; Forbes (Wade Shepard), "An Update on China's Largest Ghost City".
11. doc Zhengdong: 1.3M residents per district government (2023); 27.5% single-year population growth (2019–20); ~10× property appreciation: Wikipedia: Zhengdong New Area; Bloomberg, "Chinese Ghost Cities Are Finally Stirring to Life," Sept. 2021.
12. doc Mechanism of Zhengdong's turn — Dec. 2013 metro connection, Foxconn's Zhengzhou iPhone plant (~half of world iPhone assembly), government-office relocation: Bloomberg, ibid.; APA California Northern reprint.
13. doc The revisit literature on ghost cities filling with time: Wade Shepard's Forbes reporting; Vagabond Journey, "5 Chinese Ghost Cities that Came Alive".
14. doc China urban share 19.4% (1980) → ~67% (2024); ~800M people urbanized: Statista, urbanization in China; World Bank, urban population % — China; World Bank "Urban China" report.
15. doc Shenzhen: urban core <60,000 (1979); ~17.99M residents (end-2024); GDP ¥3.87T (~$530B, 2025); SEZ mechanism: Hinrich Foundation, "Shenzhen: From Fishing Village to Global Metropolis"; CEIC, Shenzhen GDP; Shenzhen Government Online; Demographics of Shenzhen. Norway-GDP comparison: interp, from IMF-order-of-magnitude figures.
16. doc Saudi urbanization ~84–85% (85.17% in 2024), among the region's highest: Trading Economics / World Bank series; UN-Habitat Saudi Arabia country profile. "Surfboard on a flat sea": interp.
17. doc Forest City, Johor: planned 700,000; ~9,000 residents late 2025; 15–30% occupancy estimates (2026); ~$100B project: Prospect Journal, "Empty Developments, Empty Promises"; PropCashflow investment review, 2026; Resident, Apr. 2025.
18. doc Xiong'an: ¥860B (~$119B) invested by Feb. 2025 (state figure); SOE headquarters relocations ordered (China Satellite Network Group 2024; Sinochem, China Huaneng Oct. 2025); 2026 "ghost city" coverage and sub-30% occupancy estimates: Xinhua, Apr. 2025; Wikipedia: Xiong'an; Vision Times, Mar. 2026; Brookings, "A brave new world: Xi's Xiong'an".
19. doc Pre-designation population of Xiong'an's three counties (~1M): Wikipedia: Xiong'an; occupancy interpretation: interp.
20. doc KAEC: launched 2005, master-planned for 2M residents and ~1M jobs; ~10,000 residents by 2024; ~5,126 in 2022; ~$100B ambition: Interesting Engineering, "Why Saudi's $100-billion King Abdullah city struggles to impress"; Wikipedia: King Abdullah Economic City.
21. doc McKinsey earning >$130M/yr from NEOM while assisting the audit of scope it helped design: E&T; Futurism. Characterization of the incentive loop: interp.

Interpretive claims owned by the author (no document asserts them; the documents above constrain them): the sequence thesis ("demand first, spectacle last"); the portfolio-vs-flagship account of error correction; the reading of NEOM's fraud findings as symptom rather than cause; the KAEC-as-ignored-tuition argument; the counterfactual demand-following NEOM. These are falsifiable — each names the evidence that would break it (a filled Forest City, a thriving organically-tenanted Xiong'an, a KAEC at plan, a conjured-demand city anywhere).

Method colophon

Research discipline for this essay followed a register-then-measure protocol on a Symbia durable stack: six falsifiable predictions were committed to a signed catalog (map-predictions-neom-china-article-2026-08-19) before any source was consulted, including one prediction registered with the expectation it would break. Verdicts: five held; the registered-to-break prediction (that Xiong'an would show Shenzhen-like organic success) broke as expected, and is reported above as broken (§ The control group). Verdicts and the full source registry are recorded at map-verdicts-neom-china-article-2026-08-19; the session was sealed after writing. Ordering claims rest on the chain, not on the author's memory.

Provenance pack

Everything this work rests on, retrieved by an agent given only the artifact index and asked to attempt the stranger-verification the method claims to invite. It found a P0 defect. The pack contains what survived, what did not, and what the audit got wrong — it took three passes to get right.

sha256 a4bf65ce6322b3a14071aab27b4b87ab52153f85d87d3efb397789f7c6476cca

The ordering, on timestamps

Predictions precede the essay file by five minutes and three seconds.

Still open

artifactsDigest reproduces, but trivially — the artifact list is empty. bodiesDigest did not reproduce under three canonicalizations. Every trace entry carries its own checksum and ed25519: signature and the bundle ships its public key, so the seal is verifiable in principle; a stranger cannot finish the job until the canonicalization is published. That is the next fix.

None of this makes the essay true. Every source in it could be wrong and the record would look identical. What the pack shows is the order things happened in, and where that order rests on a chain rather than on someone's word.