The Ghost City That Never Was
How the empty streets that made Ordos infamous became the asset that saved it
Ordos is not a ghost town.
The world called it that when the coal price crashed and seventy percent of Kangbashi New District went dark. Per capita GDP had doubled Shanghai’s. A futuristic city for a million residents, built on a commodity boom. Then the price turned. The lights stayed off. The cameras came. China’s ghost town became a global symbol of debt fueled hubris.
The resource curse says: easy wealth breeds brittleness. When the commodity dies, the city dies with it.
Ordos did not die. It repurposed the vacancy.
The standard playbook for a bust follows predictable stages: demolish, subsidize low value occupancy, or wait for the next commodity cycle. Ordos did none of these. It took its empty roads; designed for commuters who never arrived and asked a different question: who needs exactly what we have too much of?
The answer was autonomous vehicles.
In the mid 2010s, AV developers hit a wall. Testing in Beijing, Shanghai, Shenzhen was impossible. Too much traffic. Too much liability. Too many pedestrians. No city would offer empty roads at scale. No city could.
Ordos could. It had 355 square kilometers of paved grid, zero congestion, full fiber connectivity and a local government that had already been ridiculed which meant it had nothing left to lose. It issued the permits, waived the liability and rented the testbed to Baidu.
Today, the largest AV fleet in China operates from Ordos. The coal financed the roads. The empty roads became the asset. The asset attracted the industry that outlives the coal.
This is not a story about resilience. It is a story about reclassification.
The crash did not create the vacancy. The vacancy existed the moment Kangbashi was paved. The crash revealed it. Revelation is not catastrophe. It is inventory. And inventory, properly audited, becomes an asset no healthy city can replicate.
Zhangjiajie solved for capture. Ordos solves for Ghost Equity.
Two problems. Two principles.
Now, lets audit the boom that built the ghost.
THE BOOM LOGIC
What created the surplus
The Coal Mandate
Ordos sits on 201.7 billion tons of proven coal reserves; one sixth of China’s total.1 The seams are shallow, low in sulfur and ash and located directly under land no one was using for anything else.2 Natural gas sits beneath the same ground: 801.7 billion cubic meters proven, one-third of China’s total.3
In 2001, the State Council upgraded Ordos from a league (盟) to a prefecture level city (地级市).4 This is not administrative trivia. Prefecture status gave Ordos independent fiscal authority, municipal borrowing capacity and political signaling to investors. The timing was precise. China entered its most intense coal burning decade. From 2000 to 2010, national coal consumption nearly tripled, from 701.3 million tons to 1,743.4 million tons.5 Ordos was the closest shovel to the fastest growing fire.
Coal output exploded. By 2006, Ordos became China’s first 100-million-ton coal base.6
GDP followed. In 2010, Ordos ranked first in China for GDP per capita.7 By 2011, per capita GDP reached ¥163,012, double Shanghai’s level.8 The local government’s annual fiscal revenue exceeded ¥50 billion at peak.9
The Kangbashi Decision
The money had to go somewhere. The old urban center, Dongsheng, was constrained by terrain and fragmented property rights. The government chose greenfield: Kangbashi New District designed for one million residents.10 Wide roads. Central heating and cooling. Fiber backbone. Monumental public buildings.
Initial investment exceeded $1 billion in early 2000s money.11
The Overbuild as Strategy
This was not incompetence. It was speculative overbuild as strategy.
The logic: coal prices stay high → fiscal revenue stays high → population arrives → infrastructure fills. The vacancy was not a bug. It was future capacity, built ahead of demand because the capital was available now and might not be later.12
The bet lost. But the infrastructure remained.
The surplus was not an accident. It was a speculative bet that coal would stay high long enough for organic fill to catch up.
THE CRASH
Why surplus became liability
The Price Collapse
The coal price that built Kangbashi did not stay high.
In 2011, thermal coal peaked at approximately ¥850–¥900 per ton.13 By 2015, it had collapsed to ¥350–¥400 per ton a drop of more than 55 percent.14 This was not a standard cyclical downturn. It was structural. China’s economy, which had grown at double digit rates for two decades, slowed to 6–7 percent. More consequentially, the central government announced a deliberate pivot away from coal. Smog choked cities, international climate pressure and a strategic calculation that renewables and natural gas would define the next era made coal a politically toxic asset.15
Ordos had built its entire fiscal architecture on the assumption that coal would remain the foundation of Chinese energy for the foreseeable future. That assumption collapsed in less than four years. The city was not diversified. It was not hedged. It was leveraged to a single price and that price moved against it with the force of national policy.
The Fiscal Air Left the Room
Ordos’s local government had structured its finances around perpetual growth. Bonds had been issued against future revenue. Land had been acquired and developed based on projected population trajectories. Kangbashi’s roads, fiber backbone, central utilities and monumental public buildings were already paid for but the carrying costs did not disappear.
Annual fiscal revenue, which had peaked above ¥50 billion, contracted sharply.16 The city could no longer subsidize the vacancy it had built as future capacity. Maintenance costs for 355 square kilometers of urban infrastructure remained. Debt service on bonds issued during the boom remained. The population that was supposed to fill Kangbashi had not arrived and now the revenue to wait for them had evaporated.
This is the hidden mechanism of the resource curse that most accounts miss. The curse is not that the money runs out. The curse is that the money runs out after you have built the permanent infrastructure that requires continuous funding. Ordos was not broke. It was broke with a mortgage on a million person city that had no tenants.
The Occupancy Trap
Kangbashi was designed for one million residents.17 By 2015, actual population was estimated at 300,000 most of whom were government employees mandated to relocate from Dongsheng as part of a political directive to legitimize the new district.18
Commercial vacancy exceeded 70 percent. Residential vacancy was similar. Streetlights operated at 5 percent of designed utilization. The central heating plant that was built to warm a million people ran at a fraction of capacity, bleeding money every winter.
Global media arrived. CNN, the BBC and the financial press ran the same photographs: wide empty boulevards, monumental civic buildings with no one entering or leaving, a city without citizens.19 The label stuck: China’s ghost town. Kangbashi became a global symbol of debt fueled hubris, a cautionary tale told whenever anyone proposed building ahead of demand.
But the media narrative missed something critical. The vacancy was real. But it was not random. It was a specific kind of vacancy: new infrastructure, fully permitted, zero congestion, no legacy systems to tear out and a local government that had already been humiliated. The shame, as it would turn out, was an asset.
The Standard Options (All Bad)
The conventional post bust playbook offered three paths, none of which Ordos could afford.
Demolish. Tear down vacant buildings, consolidate population into a smaller footprint. This was politically impossible. It would admit total failure and waste the sunk cost of a decade of infrastructure investment. It would also require the local government to explain to Beijing why a centrally approved development zone had been abandoned. No mayor survives that conversation.
Subsidize low value occupancy. Offer cheap rent to whoever would come. This was the path of least resistance and Ordos rejected it. The reason is instructive: once you fill prime infrastructure with low value users (storage, informal retail, marginal industry), you cannot later upgrade them. The district would become a permanent slum of surplus capacity. The ghost town was embarrassing. A slum would be permanent.
Wait for the next coal cycle. Passive. Fatalistic. Coal might never return to peak prices, and even if it did, the structural shift toward renewables meant the next cycle would be weaker. Waiting was not a strategy. It was a slow liquidation of the city’s remaining fiscal capacity.
Ordos did none of these. It chose a fourth path that no one had written a playbook for.
What the Crash Revealed
The crash did not create the vacancy. The vacancy existed the moment Kangbashi was paved in the mid 2000s. What the crash revealed was that the original use case; commuters filling the district organically was never arriving.
This is the diagnostic shift that separates Ordos from every other post boom ghost story. Most cities see vacancy as failure. They demolish, or subsidize low value occupancy or wait. Ordos looked at its empty roads and asked a different question: who needs exactly what we have too much of?
The bet had lost. But the infrastructure remained. And infrastructure, even empty, has specifications: lane width, surface quality, connectivity, permitting status, absence of pedestrians. Those specifications were not liabilities. They were inventory. And inventory, properly audited, becomes an asset no healthy city can replicate.
The crash didn’t create the surplus. It revealed it. The surplus existed the moment Kangbashi was paved. The crash just removed the illusion that commuters were coming.
THE LIABILITY INVENTORY
What Ordos actually had
The Asset Hidden Inside the Liability
Most cities looked at Kangbashi and saw failure. Ordos looked at Kangbashi and saw specifications.
The distinction is the entire principle. Failure is a judgment. Specifications are a fact. The district had been built to a certain standard: 355 square kilometers of paved road grid, designed for high speed automated traffic flow. Fiber optic backbone to every building, installed because the boom had money to bury cable. Central heating and cooling plants, oversized by design to accommodate future density. Traffic lights, streetlamps, drainage systems, all new, all functional, all operating at a fraction of their designed capacity.20
The vacancy was not uniform. It was not decay. It was surplus infrastructure in mint condition.
The local government had something no other Chinese city could offer: a full scale urban environment with zero congestion, zero pedestrian interference, zero legacy systems to retrofit and full regulatory authority to issue permits for experimental use. The roads were not empty because the city had failed. They were empty because the intended users had never arrived. That emptiness was the feature.
Why No Other City Could Compete
Healthy cities cannot offer what Ordos had. Beijing’s roads are saturated. Shanghai’s streets have pedestrians, bikes, delivery scooters and a thousand variables that make autonomous vehicle testing impossible. Shenzhen is dense. Guangzhou is dense. Even second tier cities have traffic.21
To build an AV testbed from scratch, a city would need to: acquire land (years of litigation), pave roads (billions in capital), install fiber (more billions), issue permits (political risk) and then keep people off the roads while testing occurs. No mayor would approve it. No electorate would tolerate it. No bank would finance it.
Ordos had already done all of this. The land was acquired. The roads were paved. The fiber was buried. The permits could be issued because the local government had total authority and no one living in Kangbashi to complain about the noise. The emptiness was not a bug. It was a pre-built greenfield that had been accidentally optimized for exactly the industry that was about to need it.
The only cost Ordos could not replicate was the political cost of deliberate overbuild. No city would build a ghost town on purpose. Ordos had built one by accident. That accident was a monopoly.
The Stigma as Immunity
There is a second asset hidden inside the ghost town narrative that no economic model captures: the freedom that comes from having already been humiliated.
By 2015, Ordos had been the subject of hundreds of ghost town articles. CNN flew drones over Kangbashi. The BBC ran the empty boulevard footage. The financial press used Ordos as a symbol of everything wrong with China’s debt fueled growth model.22 The city had no reputation left to protect.
This matters because experimentation requires failure tolerance. If Baidu’s autonomous vehicles crashed into a traffic light in Beijing, the political fallout would be immediate. Lawsuits. Regulatory inquiries. Headlines. In Ordos, the headline would be: Robot car crashes in ghost town. The marginal damage to the city’s reputation was zero. It had already been called a failure.
The local government understood this implicitly. They had nothing to lose. That made them willing to issue liability waivers, fast-track permits, and accept risks that no healthy city would accept. The stigma was not a cost. It was a subsidy.
The Inventory Audit
By 2015–2016, the conditions were undeniable. Kangbashi had been built in three stages from 2004 to 2015.23 Over $1 billion had been invested.24 The standard playbook had three options: demolish, subsidize low value occupancy or wait. Ordos did none of these.
Instead, the city began a series of quiet actions that reveal an implicit inventory audit. The local government did not tear down the empty buildings. It did not auction off the vacant land to the lowest bidder. It maintained the infrastructure; the roads, the fiber, the central utilities at a level that suggested it was waiting for something specific.
The city asked three questions:
What do we have too much of? Roads. 355 square kilometers of paved grid.25 Fiber backbone. Central heating and cooling. Traffic lights, streetlamps, drainage all new, all functional, all operating at a fraction of designed capacity.
Who needs exactly that? In 2015, the answer was unclear. By 2017, it was Baidu and the autonomous vehicle industry. AV developers had reached a point where simulation was no longer sufficient. They needed real roads, real conditions, but without the chaos of real cities. Ordos was the only place on earth offering a full scale urban grid with zero entropy.
What is the minimum viable transaction? Ordos did not need to own the AV IP. It did not need to manufacture vehicles. It did not need to capture value through direct tolls. The minimum viable transaction was: rent the testbed. Baidu pays in data center construction, talent residency, secondary investment and the narrative shift from ghost town to future town.
The inventory audit revealed that the city’s greatest liability; empty infrastructure was also its only non-replicable asset. The crash had not destroyed Ordos. It had stripped away the false use case, revealing the true specifications underneath. You cannot deliberately build a ghost town. Ordos accidentally built a monopoly.
Key Diagnostic: The surplus was not a problem to solve. It was inventory to be reclassified.
THE PIVOT
How liability became asset
The Customer Discovers the Asset
In October 2017, Baidu signed a strategic cooperation agreement with BAIC Motor to mass produce Level 3 and Level 4 autonomous vehicles by 2021.26 In the same month, Baidu signed with King Long to mass produce autonomous buses by July 2018.27 In October 2018, Ford announced a joint AV testing project with Baidu in Beijing.28 Baidu was building an ecosystem, and it needed real world validation data at scale.
Beijing could not provide that data. Too much traffic. Too much liability. Too many pedestrians. But Ordos had roads; 355 square kilometers of empty, paved, fiber connected roads, fully permitted and ready for use.
According to reporting from 2018, Baidu began testing autonomous vehicles in Ordos that year.29 The transaction was not announced with a ribbon cutting. It was quiet. A few vehicles. A few permits. A local government that had nothing to lose saying yes to everything.
The asset did not move. The customer moved toward the asset. And the only reason Baidu came was that no other city could offer what Ordos had: empty roads at scale, with a government willing to waive liability and fast track permits and zero reputational risk for failure.
The pivot was not a marketing campaign. It was a convergence of two trajectories; a city that had overbuilt during a commodity boom and an industry that had outgrown simulation but could not afford to build its own test environment. The convergence was not inevitable. It was improvised.
The Tripartite Stack
Ordos did not simply open its roads and hope. It constructed a governance structure that made the testbed work for all parties.
Local Government as Platform Provider and Regulator. The Ordos municipal government established the Smart Connected Vehicle Testing and Application Working Group, led by the Municipal People’s Government and including the Bureau of Industry and Information Technology, Public Security Bureau, Transportation Bureau) and Energy Bureau.30 This working group had the authority to issue testing permits, issue temporary license plates, manage traffic violations and accidents, and open testing roads by decree .
The Third Party Management Mechanism. The working group commissioned a third-party management organization to handle daily operations: accept applications, organize an expert committee, conduct qualification certification of test subjects, safety officers and vehicles and monitor operational data. The Expert Review Committee included experts from AV companies, research institutes and enterprises . This created arm’s length legitimacy without requiring a specific academy anchor.
Baidu as Operational Anchor. Baidu brought the capital, the engineering talent and the brand. Its Apollo autonomous driving platform became the operating system for the Ordos testbed. By 2018, Reuters reported Baidu testing vehicles in Ordos.31
The CAS involvement that does exist came later (2023) and was specific to mining vehicles, not the Kangbashi road testbed.32
This tripartite stack; government, academy, private anchor is not unique to Ordos. But Ordos executed it with unusual speed because the ghost town had eliminated the usual friction of stakeholder negotiation. There was no one to object.
What Ordos Gained
The city did not sell the testbed. It rented it. The benefits were not direct toll revenue. They were soft capture, harder to measure, but harder to lose.
Data Center Construction. Baidu and other AV developers built computing infrastructure in Ordos to support testing operations. Those data centers remain after the testing ends. The coal boom built the roads. The AV boom built the servers.
Talent Residency. Engineers relocated to Ordos, even temporarily. They needed housing, food, transportation and local services. The ghost town gained residents, not a million, but a specialized class of workers whose presence signaled that something was happening.
Secondary Investment. Suppliers followed Baidu. Sensor manufacturers, mapping companies, logistics firms the ecosystem that builds around an anchor tenant. The city did not recruit them. They came because Baidu was there.
Narrative Shift. The global media narrative flipped. Ghost town became AV testbed. The city that had been a symbol of failure became a symbol of reinvention. This narrative shift attracted more investment, more talent, and more policy attention from Beijing.
Regulatory Precedent. By hosting the first large-scale AV testbed, Ordos wrote the rules that other cities would later copy. The city captured not just economic value but institutional value the authority to define standards, certify safety officers, and approve test routes. That authority is a tollgate that does not depend on Baidu’s continued presence.
The city did not need to own the AV IP. It owned the testbed. The testbed is immobile. The IP rents the testbed. And if Baidu leaves, another developer will arrive. The vacancy that was once a liability is now a perpetually rentable asset.
GHOST EQUITY: THE OTHER APPLICATIONS
Ordos did not apply the Ghost Equity Principle once. It applied it at least four times. The roads were the most visible. The schools, universities and house ticket were equally strategic.
Each application follows the same diagnostic sequence: identify a liability, audit its latent specifications, reclassify the asset and rent it to a captive demand source. The pattern is not accidental. It is a protocol.
The School as Population Acquisition Device
Ordos had empty apartments. The standard playbook said: lower prices, attract any buyer. Ordos did something else. It moved the schools.
In 2007, the city partnered with Beijing Normal University to establish the Beijing Normal University Ordos Affiliated School, Kangbashi’s first school, opened September 1, 2007 with 141 students and 46 teachers recruited nationally.33 The city spent 77.22 million yuan on educational infrastructure that year.
In 2010, the city made its boldest move. Ordos No. 1 High School the city’s top tier high school, relocated entirely to Kangbashi. Sixty classes. 3,300 students. 270 staff. Not a branch. The whole school.34
The relocation was not about education. It was about bodies. As Japanese media reported: the Ordos municipal government relocated some of the city’s top schools to Kangbashi; tiger parents followed and housing prices soared.35
The mechanism was precise: strict school district zoning meant that enrollment required residence in the corresponding zone. Parents did not buy apartments. They bought access to schools.
The price movement tells the story. Kangbashi housing prices collapsed from 8,000 yuan per square meter to between 3,000 and 5,000 yuan after the crash. By 2021, prices in the central district had rebounded to 15,000 yuan per square meter. Ordos locals generally attribute Kangbashi’s housing price increase to the improvement in educational resources.36
By 2024, the model had scaled. The China Basic Education High Quality Development Index ranked Kangbashi fifth nationally, first in western China. Every compulsory education school in the district met quality standards.37
The same concrete boxes. A different classification.
When covering Hengshui (City Ten), we documented how the gaokao reshaped a city as an emergent outcome. Ordos is different. Here, education was not a byproduct. It was a lever. The city moved its top high school into the ghost town deliberately, not because the system demanded it, but because the vacancy demanded a fix. Hengshui's transformation happened to it. Ordos's transformation was engineered.
The University as Talent Acquisition Device
The school playbook worked. Ordos applied it again, one level up.
In 2008, the city partnered with Inner Mongolia University to establish Inner Mongolia University Ordos College ending Ordos’s history of having no university. The city built the campus. Inner Mongolia University sent the faculty.38
In 2009, Ordos Vocational and Technical College landed in Kangbashi. By 2011, a full education pipeline had been built: a complete education system from kindergarten to university had taken initial shape in Kangbashi.
But Ordos did not stop at local partnerships. The city began recruiting directly from China’s top universities.
By 2025, Kangbashi had attracted graduates from Tsinghua University, Zhejiang University, Fudan University, Wuhan University and Peking University through the Warm City Invitation) recruitment program.39 The city’s talent strengthening district strategy offered state owned enterprise recruitment with tiered subsidies of 10,000 yuan for undergraduates, 20,000 for masters, 30,000 for doctorates.40
The city had moved beyond importing universities. It was now importing the students directly.
By 2025, Kangbashi had cumulatively introduced 46 high level talents from top tier universities, with an additional 34 experienced teachers recruited.41 The district had issued over 23.62 million yuan in talent subsidies, covering housing, rental and living expenses for 453 individuals.42
The Ordos Eco-Environmental Vocational College continued the pattern, announcing 64 new faculty positions in 2026 with doctorates included in government staffing and salaries equivalent to civil service Grade 6 and 8 management positions.43
The university branches and talent recruitment were not academic assets. They were population acquisition devices. Empty buildings require people. People follow education and employment. Ordos built the supply. The demand followed.
The House Ticket as Inventory Reclassification
Ordos had a third liability demolition compensation. Traditional method: pay cash. The cash leaves the city. Ordos needed the cash to stay inside its empty buildings.
In 2016, Ordos’s Dongsheng District issued the Ordos Dongsheng District Housing Exchange Certificate known as the house ticket.44
The instrument was simple. received not cash, but a certificate redeemable only for housing inventory in the district. The ticket came in denominations from 1 to 100 square meters, each marked with zone type, price per square meter and expiration date.
The Housing ticket could be:
Split across multiple properties
Transferred to other buyers
Traded between private parties at negotiated discounts
But it could not leave Ordos.
The Fiscal Engineering
The Dongsheng District Housing and Land Exchange Certificate Management and Settlement Center was established to issue, settle and cancel the certificates.
The economics were precise. Demolition households’ old properties were assessed. The Housing ticket was issued at a fraction of peak housing prices. The lowest zone category converted to 3,822 yuan per square meter, just a fraction of Ordos’s peak housing prices).45
But for Demolition households who had lived in old houses for decades, the Housing ticket was an upgrade. For the city, the Housing ticket was a closed loop.
By August 2016, the Dongsheng District had issued approximately 262 million yuan worth of Housing tickets (65,300 square meters), with 586 million yuan worth (14,400 square meters) already redeemed.
The city added additional incentives to keep the loop tight. Rural residents moving to the city received 200 yuan per square meter in subsidies. Poverty alleviation relocation recipients received the same. Rural dangerous old house demolition recipients received 21,500 yuan per household.46
By 2025, the inventory reduction strategy had absorbed 561 units of unsold housing inventory, though the city still faced challenges with light on rates in some districts remaining only half occupancy.47
Every yuan of Demolition compensation that would have leaked to other cities was forced to circulate inside Ordos’s empty buildings.
The Pattern
Four different liabilities. Four different reclassifications. One principle.
Ordos did not build anything new in any of these applications. It relabeled what it already had. The roads were never for commuters. The apartments were never for residents. The Demolition compensation was never for spending elsewhere. The education system was never just about education, it was about importing population to fill empty buildings.
Each application required the same discipline: do not demolish, do not subsidize low value use, do not wait passively. Audit the specifications. Find the asymmetric need. Maintain the asset. Rent it.
The crash did not destroy Ordos. It revealed inventory. And inventory, properly reclassified, becomes a monopoly, whether that inventory is asphalt, concrete, or the human desire for a better future.
THE GHOST EQUITY MECHANISM
How it works
Defining Ghost Equity
Ghost Equity is the value trapped inside an asset that appears worthless because its intended use has failed. The unlock is not renovation or demolition. It is reclassification.
Ordos’s roads were built for commuters. Commuters never came. But autonomous vehicles do not need commuters. They need asphalt and absence. The asset’s value was not in its performance; moving people. It was in its latent specifications: width, surface quality, connectivity, emptiness, permitting status.
Most cities look at failed infrastructure and see sunk cost. Ordos looked at failed infrastructure and saw inventory. The difference is not semantic. Sunk cost is a judgment about the past. Inventory is a fact about the present. Inventory can be sold, rented, or reclassified. Sunk cost can only be written off.
The principle applies to any asset whose original use case has collapsed: empty office towers after a remote work shift, abandoned shopping malls after e-commerce, surplus cold storage after a supply chain reroutes. The question is not what was this for? The question is what can this do?
Why the Asset Cannot Be Replicated
No other city can deliberately build a ghost town to attract AV testing.
The political cost is prohibitive. What mayor campaigns on we will build empty roads and wait for an industry to discover them? The financial cost is prohibitive. No bank lends for intentional vacancy. The timing cost is prohibitive. The boom-bust sequence that created Ordos took a decade and a commodity super cycle. It cannot be compressed.
Ordos has a negative monopoly. It owns the only large scale, pre-built, fully permitted, zero-traffic urban grid on Earth. The vacancy was not a strategy. It was an accident. But accidents, properly audited, become assets.
The Ghost Equity Protocol
The principle extracts to five steps that any post crash city can test.
Overbuild first, or inherit a crash’s leftovers. Surplus must exist before it can be reclassified. Ordos built Kangbashi during the coal boom. Other cities inherit surplus from failed industrial zones, abandoned military bases, or collapsed real estate projects.
Audit your liabilities for latent specifications. Do not ask what went wrong? Ask what is actually here? Lane width. Surface quality. Connectivity. Permitting status. Absence of friction. These are not judgments. They are facts.
Wait for an industry to emerge that needs exactly what you have too much of. This is the hardest step because it requires patience. Ordos waited from 2012 to 2017. The industry came. It might not come for every city. But if you demolish, you guarantee it never comes.
Do not renovate for the original purpose. Reclassify for a new one. The roads were never for commuters. They were always for something else. The crash just cleared the mistaken assumption.
Rent the asset. Never sell it. The testbed is immobile. That is your only true monopoly. Sell access, not ownership.
When the Principle Fails
Ghost Equity is not a universal solution. It fails under three conditions.
The surplus has no unique specifications. If your vacant infrastructure is generic; poorly built, poorly located, poorly connected, no industry will rent it. Ordos’s roads were unusually well built because the coal boom financed high specifications. Generic vacancy is just vacancy.
No emerging industry needs your constraint. Ordos was lucky. The autonomous vehicle industry arrived at exactly the moment the city had empty roads. If the crash had happened five years earlier or five years later, the convergence might not have occurred. Ghost equity requires alignment. Alignment cannot be forced.
The local government lacks regulatory authority. Ordos could issue permits because it was a prefecture level city with independent fiscal and regulatory power. A county or township with no authority to waive liability or designate test routes cannot execute the pivot.
Why This Matters for You (The Reader)
You do not need to have built a ghost town to use this principle. You need to be holding an asset whose original use case has failed, and to have the patience to wait for the right renter.
The relevant question is not How do I become Ordos? The relevant question is: What do I have too much of, and who needs exactly that?
Ask yourself three questions.
First, what are you holding that everyone else calls a failure? Ordos held empty roads. The world called it a ghost town. The label was not wrong. It was incomplete. The roads were empty. That was a fact. The judgment of failure was an interpretation. Separate the fact from the judgment. The fact is inventory. The judgment is noise.
Second, what are the latent specifications of your asset that no one else can offer? Ordos’s roads were empty. But they were also new, paved, fiber connected, fully permitted, and located in a jurisdiction with nothing left to lose. That combination could not be replicated. Audit your own asset. Not its intended purpose. Its actual specifications. Lane width. Permitting status. Absence of friction. Location. Connectivity. These are facts. List them.
Third, can you afford to wait? Ordos waited five years from the crash to Baidu’s arrival. The city did not demolish. It did not subsidize low value occupancy. It maintained the infrastructure and waited. If you cannot wait; if carrying costs will force demolition or fire sale, ghost equity is not available to you. The principle requires patience. Patience is not a virtue. It is a structural requirement.
The Ghost Equity Principle is not a story about resilience. It is a diagnostic framework. Apply it to your own failed asset, your own surplus, your own constraint. Audit the specifications. Find the asymmetric need. Maintain the asset. Wait.
Then the right renter arrives. And the ghost town becomes a testbed.
The Roads Were Never Empty
Ordos did not solve the resource curse. It sidestepped it.
The coal is still there. The price could crash again. Baidu could leave. The testbed could go dark. The ghost town could return.
That is not a failure of the principle. It is the principle.
Ghost equity is not a permanent escape from volatility. It is a bridge. It buys time. It converts a liability into a rental income stream. But the liability remains. The vacancy remains. The dependence on a single tenant remains.
Ordos is not a success story. It is a diagnostic. It proves that surplus can be reclassified. It does not prove that reclassification lasts forever.
The question the reader takes home is not how do I build a ghost town? It is: If my renter leaves tomorrow, what do I own that the next renter will also need?
Ordos’s answer: empty roads. The renter changed. The asset did not.
What is your empty road?
That is the Ghost Equity Principle. You do not need a beautiful landscape. You need the courage to not demolish your ruins and the curiosity to discover who might rent them.
NEXT WEEK: ZHENJIANG
Ordos held empty roads until the right renter arrived. That was patience.
Zhenjiang holds something else: a river, a vinegar and a history of being overlooked. Jiangsu’s smallest prefecture level city by land area, sandwiched between Nanjing and Yangzhou, has spent decades watching its neighbors capture the attention, the investment, and the talent.
But Zhenjiang has a card Ordos does not. It has a product. A specific one. A product that cannot be moved, cannot be copied and has been sitting on shelves for 180 years waiting for someone to realize that the bottle is not the business.
The question is not whether Zhenjiang has an asset. The question is whether it knows what kind of asset it is holding.
Lets look at Zhenjiangm, next week.
SOURCES
China Daily. “Ordos – Abundant Resources.” (2024).
China Daily. “Ordos – Resources.” (2026).
ibid1
BP Statistics (2016). China coal consumption data (2000–2010).
Economic Reference News (2006).
Wikipedia. “List of prefecture-level cities by GDP per capita.”
Aju News. (2010).
China Daily. “From ghost town to high-tech hub.” (2023).
World Bank Commodity Price Data (coal, Australian thermal, 2011).
World Bank Commodity Price Data (coal, Australian thermal, 2015).
Statistics of China, GDP growth rates (2010–2015).
Inner Mongolia Provincial Statistical Yearbooks.
CNN, “Inside China’s largest ghost town” (2016). https://www.cnn.com/style/article/china-ordos-ghost-town/index.html
South China Morning Post, “Ordos: the Chinese ghost town that refuses to die” (2018).
World Economic Forum, “Why China’s ‘ghost city’ of Ordos is now a hub for autonomous vehicles” (2021).
ibid19
ibid24
HKEX News: BAIC Motor Corporation Limited, Voluntary Announcement Entering into the Strategic Cooperation Framework Agreement with Baidu (October 16, 2017).
TechNode, “Baidu signs strategic partnership agreements to accelerate autonomous driving” (October 20, 2017).
Ford Media, “Ford and Baidu Announce Joint Autonomous Vehicle Testing” (October 31, 2018).
Ordos City Intelligent Connected Vehicle Testing and Demonstration Application Management Implementation Rules (Trial)
ibid29
CAS The Institute of Automation, Chinese Academy of Sciences, and other institutions jointly incubated the “CarMo” intelligent transportation robot for new energy mines, which was officially released.
Sina News (Nikkei Asian Review): “Housing prices in Kangbashi District, Ordos, are soaring; Japanese media: Relying on school relocation to save the real estate market is unsustainable.
People’s Daily Online, 2025 “Kangbashi: Talent and City Moving Towards Each Other
ibid33
ibid33
ibid33
ibid37
ibid34
Ordos Municipal People’s Government “Kangbashi District focuses on innovation and breakthroughs, and its talent pool and empowerment initiatives have achieved new progress.” (2025).
Ordos Municipal People’s Government: “Attracting Talent to a Healthy Life and Drawing a Blueprint” (2025).
The website for university talent recruitment announced that Ordos Ecological and Environmental Vocational College will recruit 64 talents in February 2026.” (2026).
People’s Daily Online Finance: “Ordos, the ‘ghost city’ after the bursting of the housing bubble: Housing voucher system to reduce inventory” (2016).
ibid44
ibid44
Problematic Housing: “Ordos, the fifteenth stop! When the trillion-yuan GDP of the ‘new energy tycoon’ collides with the decade old scars of the ‘ghost city’.” (2026).
Lile, I loved this.
What makes the piece compelling is that it refuses the usual redemption narrative. Most stories about struggling cities are framed as resilience or comeback stories. Instead, you frame Ordos as an exercise in reclassification.
I found myself thinking about how often this principle applies far beyond cities. This passage especially stood out: “Revelation is not catastrophe. It is inventory.”
What a remarkable way to think about both cities and people. Another wonderful essay, Monica
Great story how to overcome adversity