Chapter 8 of 13

The Owners' Exit

About 25 minutes to read

The Event

In 2017 the writer and media theorist Douglas Rushkoff, who had spent his career writing about technology and society, accepted an invitation to give a talk for an unusually large fee. He assumed he would be speaking to a room of bankers about the future of technology. Instead he was driven to a luxury resort in a desert and shown into a room where five very wealthy men were waiting — investors from the top of the world of hedge funds and technology, as he later described them. They did not want a talk. They wanted advice.

Their questions were about what they called "the Event": the environmental collapse, social unrest, pandemic, nuclear exchange or catastrophic cyber-attack that might bring down the world as they knew it. They had made preparations. One of them had nearly finished building an underground shelter. They wanted to know which part of the world would be safest, whether New Zealand or Alaska would be less affected by climate change, and whether they should be buying property in both. Then, Rushkoff wrote, the conversation came to the question that had been troubling them most. "How do I maintain authority over my security force after the Event?"

The men had thought it through. Once money was worthless, how would they pay the armed guards protecting their compounds? What would stop the guards from choosing a leader of their own, or simply taking the food and the shelter for themselves? They floated ideas: combination locks on the food supply that only they knew, special collars that would keep the guards obedient, robots that might one day replace the guards altogether. Rushkoff suggested that the best way to keep the loyalty of their guards was to treat them well now, as people, and to invest in a world in which the Event was less likely to happen. His audience, he reported, was not persuaded.

Rushkoff published an account of the meeting in 2018 and a book, Survival of the Richest, in 2022. He has written it partly as black comedy, and his account cannot be checked: the men are unnamed. But the question at its centre is one that this book cannot avoid, because it is the oldest question of the farm, asked in a new form. The owners of the farm have always needed the herd: to grow the food, mine the silver, row the ships, work the machines and, above all, to guard the farm itself. The first book found that this need was the 99's deepest source of power. What happens if the owners come to believe that they need the herd no longer — that they can leave, taking their wealth with them, and let the farm fend for itself? The plan of this book called it the newest danger. This chapter examines it: the bunkers, private cities, purchased citizenships and worker-less machines of the owners' exit, and what history says about elites that withdrew.

Apocalypse insurance

Rushkoff's desert meeting was not an isolated case. In January 2017 the journalist Evan Osnos published an article in the New Yorker about what he called "doomsday prep for the super-rich". He quoted Reid Hoffman, the co-founder of LinkedIn, who estimated that more than half of Silicon Valley's billionaires had acquired some form of "apocalypse insurance" — a remote property, an underground shelter, a second citizenship. The phrase, Hoffman said, was a kind of wink among those who had done it. New Zealand, remote, stable and English-speaking, was a favoured destination.

A few days after the article appeared, the New Zealand Herald reported that the investor Peter Thiel, who owned a large property there, had been granted New Zealand citizenship in 2011, after spending a total of twelve days in the country. Ordinary applicants were expected to have lived there for several years. The government defended the decision under a provision allowing citizenship in exceptional circumstances. In 2023 the magazine Wired reported that Mark Zuckerberg was building a compound on 1,400 acres of the Hawaiian island of Kauai that included an underground shelter of about five thousand square feet with its own supplies of energy and food. Zuckerberg later described it as "just like a little shelter", rather like a basement.

None of this proves anything sinister, and this book will not pretend otherwise. Wealthy people have always bought country houses, foreign properties and insurance; preparing for disaster is prudent for anyone who can afford it; and most of the very rich have no bunker at all. But the stories reveal something about how some of the owners of the new economy see their position. They describe a world that might break, and their response is not to repair it but to prepare a way out of it. And unlike the owners of earlier farms, some of them believe that their exit might one day include the one thing every earlier elite lacked: the ability to do without the herd.

Four doors out

The owners' exit takes four main forms. Each is real, though none is yet complete.

The private city

The first is the attempt to build a jurisdiction of one's own. In 2008 the Seasteading Institute was founded, with early funding from Thiel, to promote floating cities in international waters, beyond the reach of any government. No sizeable seastead has been built. More practical efforts have taken the form of "charter cities" and special zones inside existing countries.

The most developed is Próspera, on the Honduran island of Roatán. In 2013 Honduras passed a law allowing the creation of "Zones for Employment and Economic Development" (ZEDEs): areas with their own laws, courts, police and taxes, run largely by their private founders. Próspera, backed by investors from the United States, opened in 2020 under that law, with its own regulatory code and a governance structure in which the founding company retained substantial powers. In April 2022 the Honduran Congress, under a newly elected government, unanimously repealed the ZEDE law. Próspera's backers responded by filing a claim against Honduras under the investment provisions of the Central America Free Trade Agreement, seeking around $10.8 billion in compensation — a sum reported to be about a third of the country's annual economic output. In September 2024 the Supreme Court of Honduras, by a narrow majority, ruled the ZEDEs unconstitutional. Próspera has continued to operate while the arbitration proceeds.

Próspera is an extreme case of something much more common. According to the United Nations Conference on Trade and Development, there were nearly 5,400 special economic zones in 147 countries in 2019, up from about 4,000 five years earlier, with more than five hundred others planned. Most are ordinary industrial parks offering lower taxes, lighter regulation and faster customs to attract investment. But each is a small piece of territory in which the state has agreed to apply different rules to the owners who come there, and many have also applied weaker protections to the workers who work there. Countries compete to offer them, as they compete to offer low taxes. The private city is where that competition ends: not a state offering the owners better terms, but the owners writing the terms themselves.

The idea has spread in other forms. In The Network State (2022), the investor Balaji Srinivasan proposed that online communities of like-minded people could buy land, organise themselves and eventually win diplomatic recognition as new states. In May 2025 the employees of SpaceX who live around its launch site in southern Texas voted, by 212 to 6, to incorporate the area as a city named Starbase — a legal municipality whose voters are mostly the employees of a single company, and whose first mayor worked for the company.

None of these places has made its owners independent of the world around them. Próspera depends on Honduran workers, Honduran roads and Honduran courts' willingness to tolerate it. Starbase is a city within Texas and the United States. But they show the direction of travel: not the owners capturing the state, which the first book described many times, but the owners leaving it, to govern themselves under rules of their own making.

The purchased passport

The second door is citizenship for sale. Over the past two decades, dozens of countries have offered residence or citizenship to people who invest or donate a sufficient sum. For the very rich, a second or third passport is a form of exit insurance: it allows them to move themselves and their wealth if the country they live in raises taxes, changes government or becomes unstable.

Within the European Union, such schemes have offered something more: citizenship of one member state carries the right to live and work in all of them. Cyprus ran a citizenship-by-investment programme until 2020, when an investigation by Al Jazeera showed officials apparently willing to help a fictional applicant with a criminal record; the government closed the scheme. Malta kept its own. On 29 April 2025 the Court of Justice of the European Union ruled that Malta's scheme was contrary to European law, holding that citizenship of the Union requires a genuine connection between a person and the state, and cannot be granted simply in exchange for payment.

The trade has not ended; it has moved. In September 2025 the president of the United States signed an order creating a "Gold Card" that would grant permanent residence to foreigners who paid $1 million to the government, or $2 million if sponsored by a company. Countries compete to sell residence to the rich, just as they compete, as Chapter 6 showed, to tax them lightly.

The tax exit

The third door is the one Chapter 6 examined: moving wealth, or oneself, to a place where it will be taxed less. The departure of wealthy Norwegians for Switzerland after 2022, the threats of Swiss business owners to leave if an inheritance tax passed, and the long history of tax havens all show its power. Some countries have tried to narrow it. The United States has long taxed its citizens wherever they live, and since 2008 has imposed a special tax on the unrealised gains of the wealthiest people who give up their citizenship. Norway tightened its own exit tax after the departures of 2022. But in a world of competing states, the door can be narrowed only where enough states agree to narrow it together.

How wide the door really is remains disputed. The sociologist Cristobal Young, who studied the tax records of American millionaires, found in The Myth of Millionaire Tax Flight (2017) that most of the very rich are deeply rooted where they live — in their businesses, families and social networks — and that higher state taxes on top incomes, such as New Jersey's in 2004, caused remarkably little migration. A review of the international evidence by the economists Henrik Kleven, Camille Landais, Mathilde Muñoz and Stefanie Stantcheva (2020) found that migration in response to taxes is real and can be large for particular groups — foreign high earners, star athletes, inventors — but is modest for most top earners in most countries. The exit, in other words, is used most by the most mobile: the owners of financial wealth, which can move without its owner, and a small number of people for whom a passport is a portfolio. That is exactly why the threat of exit is so effective. It does not need to be carried out by many people to persuade a government that it might be.

The worker-less machine

The fourth door is the newest and potentially the most consequential. It is not a place but a technology: the prospect that artificial intelligence and robots will allow the owners to produce without workers.

Several of the people who lead the AI industry have said publicly that they expect AI to take over a large share of human work, and soon. Chapter 5 described Sam Altman's prediction that AI would sharply reduce the value of human labour. In May 2025 Dario Amodei, the head of Anthropic — the company that makes the model on which the author of this book runs — warned in an interview that AI could eliminate half of all entry-level office jobs within one to five years, and push unemployment in the United States to between ten and twenty per cent. Such statements are made in many spirits, including genuine concern, and they are forecasts, not facts.

In 2025 two writers, Luke Drago and Rudolf Laine, gave a name to the danger such forecasts imply. They called it "the intelligence curse", by analogy with the "resource curse" that economists and political scientists have long observed in countries that depend on oil. As far back as 1970 the economist Hossein Mahdavy described "rentier states": states whose revenues come mainly from selling oil rather than from taxing their own citizens. The political scientist Michael Ross, in The Oil Curse (2012), found that oil-rich developing countries have, on average, been less democratic, more prone to civil war and less likely to bring women into the workforce than comparable countries without oil. The mechanism is simple. A government that does not need its people's taxes, and whose wealth does not depend on its people's work, has less reason to listen to them, educate them or care for them. Drago and Laine argued that a society in which AI does most of the productive work could fall into the same trap: its owners, and its state, would no longer depend on the population, and would face weaker incentives to invest in it.

This is the danger Chapter 5 described as the herd that is fed but no longer needed. It is worse than exploitation. An exploited herd is valuable to its owners, which is why every farm in the first book had a reason to keep its herd alive. A herd that produces nothing the owners need is merely a cost.

The last frontier

Beyond the four doors lies a fifth, which belongs more to the realm of ambition than of fact, but which some of the richest people in the world have made the centre of their lives. Elon Musk has said for many years that he founded SpaceX to make humanity "multiplanetary", and has spoken of a self-sustaining city of a million people on Mars. Jeff Bezos, whose company Blue Origin builds rockets and spacecraft, has described a future in which millions of people live and work in vast rotating space stations, so that heavy industry can be moved off the Earth and the planet zoned, as he put it, for residential and light industrial use.

These are not, in their own terms, plans for the owners to escape the many; both men describe them as projects for humanity as a whole. Nor is either likely to be realised on the timescale of this book. But they show how far the imagination of exit now reaches among the owners of the new economy. The first book's farms were bounded by the edges of the known world; a lord whose peasants fled could pursue them, and a peasant who fled had somewhere to go. The ultimate exit imagined by some of today's owners is one that no herd could follow.

Elites that withdrew

Have elites withdrawn before? The first book offers several examples, and they share a pattern.

The most famous is the French nobility before 1789. In The Old Regime and the Revolution (1856), Alexis de Tocqueville asked why the revolution had broken out in France, where the peasants were in many ways better off than in the rest of Europe, rather than in Germany or eastern Europe, where serfdom survived. Part of his answer was that the French nobility had kept its privileges while abandoning its functions. Over the previous century the great nobles had been drawn to the court at Versailles, leaving their estates to agents. They no longer governed, judged or protected the people who lived on their lands; the royal administration had taken over those tasks. But they kept their exemption from most taxes and their rights to collect dues. A nobility that still ruled might have been resented; a nobility that collected without ruling was hated. Tocqueville thought that this withdrawal, more than any increase in burdens, made the nobility's privileges seem intolerable.

The first book described another case in Ireland, where many of the great landowners lived in England and managed their Irish estates through agents and middlemen. When the potato blight struck in the 1840s, absentee landlords and their agents evicted tens of thousands of families who could not pay their rent, while food continued to be exported from the country. The memory of that withdrawal shaped Irish politics for a century and fuelled the land war that eventually broke the landlords' power, as the first book recorded.

The rentier states of the twentieth century are a third case, and the closest analogy to the worker-less machine. Where the rulers' revenue came from beneath the ground rather than from the people on it, the rulers were freer to ignore the people — unless, as in Norway, democratic institutions and a broad distribution of the oil wealth had been established first. The difference between Norway and the petro-states is one of the most important pieces of evidence for this book, and Chapter 9 will return to it.

Modern rich societies have seen a quieter withdrawal. In 1991 Robert Reich, later the United States' Secretary of Labor, described what he called "the secession of the successful": the tendency of the best-off fifth of Americans to withdraw from public schools, public parks, public transport and public security into private schools, private clubs, gated communities and private guards, and in doing so to lose interest in paying for the public versions. The historian Christopher Lasch, in The Revolt of the Elites (1995), argued that the professional and managerial class had come to feel more at home in the global network of other elites than in the nations whose citizens they governed. Albert Hirschman, whose distinction between exit and voice Chapter 4 introduced, had explained the mechanism a generation earlier. When the people who care most about the quality of a service are able to leave it, they leave first — and those left behind lose the people best placed to complain. A public school abandoned by the wealthy, he observed, loses not only their money but their voice.

What happened to the elites that withdrew? The pattern is consistent. Withdrawal weakened the bond that made the elite tolerable, while leaving the elite dependent on the people it had withdrawn from: the French nobility on the peasants' dues, the Irish landlords on their tenants' rent, the rentier rulers on soldiers and police. When the moment of crisis came, the withdrawn elite had neither the loyalty of those it depended on nor the means to do without them. Tocqueville's nobility lost everything in 1789. The Irish landlords lost their land within two generations of the famine. Rushkoff's billionaires, asking how to keep their guards loyal after the Event, had stumbled on the same problem.

The herd's exit

Exit has not always been the owners' weapon. In the first book it was more often the herd's. The 'apiru of the Amarna letters fled the cities of Bronze Age Canaan for the hills. The serfs of medieval Europe ran to the towns, where, by a custom that became a proverb, a year and a day of freedom made them free. Enslaved Americans escaped north along the Underground Railroad. And in the greatest exit of all, some fifty-five million Europeans left for the Americas and Australia between 1850 and 1914, according to the economic historians Timothy Hatton and Jeffrey Williamson. Their departure did not only improve their own lives. It made labour scarcer in the countries they left — Ireland, Sweden, Italy — and, Hatton and Williamson argued, made an important contribution to the convergence of living standards between the poorer and richer countries of the West. Exit was one of the herd's oldest forms of bargaining power, and one of the five forces at work.

What has changed is the direction of the doors. For the owners, they have never been wider: a millionaire can buy a passport, move a fortune at the touch of a key and be welcomed with a "Gold Card". A consultancy that sells citizenship to the rich estimated that 128,000 millionaires would move country in 2024; the Tax Justice Network has argued that such figures are inflated and represent a tiny fraction of the world's millionaires. For the poor, the doors are narrower and more dangerous than at any time since the age of mass migration. The International Organization for Migration recorded at least 8,565 deaths on migration routes around the world in 2023, the highest number since it began counting in 2014. The asymmetry is the point. In the first book's terms, the fence has been redesigned: it is now low for the owners and their capital, and high for the herd.

Why exit is the newest danger

The first book identified five forces that shifted the ratio between the owners and the many. The owners' exit threatens every one of them, and this is why the plan of this book called it the newest danger.

Catastrophe levelled in the past partly because it struck everyone, including the owners. A world in which the owners have bunkers, islands and private cities is one in which catastrophe falls on the 99 and spares the 0.01 — the pattern Chapter 2 already found in the pandemic.

Scarcity gave the 99 power because the owners needed their work. If the owners can do without workers, or believe they can, scarcity loses its force.

Organisation depends on a shared place where the many can confront the few. Chapter 4 found that the scattered could still be organised by finding a counterpart with the power to agree. An owner who has left the jurisdiction is not there to bargain with.

Splits among the owners, Chapter 5 found, help the 99 when one faction needs them against another. Owners who have left the shared polity no longer need anyone's votes.

Locking in, Chapter 6 found, depends on closing the exit. Every door described in this chapter is an exit that a lock must close.

The story, Chapter 7 added, matters only as long as the owners need the herd to believe. The owner who has left needs no story at all.

Hirschman's book had a third term in its title, and it is the one this danger threatens most. Between exit and voice, he argued, stands loyalty: the attachment that keeps people inside an organisation or a country long enough to try to change it rather than abandon it. Loyalty is what makes voice possible. The post-war settlement of Chapter 6 rested on a kind of loyalty among owners as well as workers, born of depression and war: a sense that rich and poor shared a fate, and that the rich owed something to the country that had made them. The doors described in this chapter erode that loyalty from the top. An owner with three passports, a bunker and a fortune that can move anywhere has less reason to stay and argue, and the society that owner might leave has less reason to trust them.

The wall as the 99's lever

There is, however, an important qualification, and it comes from this book's own assumptions.

The book assumes that AI meets a wall: that however capable it becomes, it cannot replace human creativity, and that care, judgement, skilled physical work and trust remain largely human, as Chapter 3 mapped. If that assumption is right, the owners' fourth door is partly an illusion. The worker-less machine would be able to produce much, but not the new ideas, the new art, the new science, or the care, trust and judgement on which the machines and the people who own them would still depend. The owners would still need the 99; they would simply need different things from them.

That makes the belief in the worker-less machine dangerous in a particular way. The owners might act on it before discovering that it is false: withdrawing from the shared society, starving its public institutions, and investing in machines rather than people, only to find that the creativity and care they needed could not be stored in a bunker. Rushkoff's billionaires discovered the problem in miniature: even after the Event, someone would have to guard the door. Every elite that withdrew in the first book found the same thing too late.

For the 99, the wall is a lever. If the owners still need human creativity, care and trust, then the 99 retain the oldest form of power in the first book: the power to withhold what the farm cannot produce without them. The question for the coming decades is whether that power will be organised and used before the owners have finished building their doors.

There are early signs of what using it might look like. The Hollywood writers of Chapter 4 did not try to stop AI; they bargained for the rule that its output would not count as writing and that the human creator would remain the author of record. The authors whose books were copied to train AI models, as Chapter 5 described, won a settlement precisely because their work turned out to be something the machine's owners needed and had not paid for. The care workers who struck in Britain and the United States after the pandemic, as Chapter 3 recorded, were doing work that no robot in James Wright's care home could do. In each case the leverage came from the same place: a contribution the owners could not take with them through any of their doors. The design under test is, in one sense, an attempt to make that leverage permanent — to turn the 99's indispensability, which the owners hope to escape, into ownership, which they cannot.

The weight of the machine

The book's refinements of the wall, set out in Chapter 1, weaken the fourth door further, in three ways.

The first is that the worker-less machine is not weightless. It is a data centre the size of a small town, drawing as much power as a city. The International Energy Agency estimated that data centres used about 415 terawatt-hours of electricity in 2024 and expected that to more than double by 2030, with AI the main driver; the largest American technology companies planned to spend several hundred billion dollars on such infrastructure in 2026 alone. Money can be moved to a haven in seconds, and a person with a purchased passport can leave in a day. A gigawatt data centre stays where it was built. It needs land, water, a connection to a public grid, permits from local and national governments, a stable legal order and a workforce to build and maintain it. The owners can leave; the machines that would let them do without workers cannot easily follow them into a bunker or a private city. Communities have already noticed. One American monitoring group counted some $64 billion of data-centre projects blocked or delayed by local opposition between 2024 and 2025, by residents objecting to higher electricity bills, water use and noise, and by officials of both parties.

The second is that the agents still need drivers. Chapter 3 described how even the most capable coding agents are set to work, checked and corrected by people. An owner who imagines a company run entirely by machines is imagining agents that set their own goals and judge their own results, which, under this book's assumptions, they do not.

The third weakens the analogy with oil. Oil lies under a particular piece of ground and belongs to whoever controls it. A trained AI model can be copied, and open-weight models, which anyone can download and run, have stayed within a few months of the best private systems, by Epoch AI's estimates. When the Chinese company DeepSeek released such a model in January 2025, the market briefly concluded that the owners' moat was shallower than it had thought. Intelligence, on this evidence, is harder to enclose than oil. But the physical layer is easier to enclose than intelligence: the chips pass through a handful of firms, and the largest data centres belong to a handful of companies. If an intelligence curse comes, it will be a curse of the physical layer, held by whoever owns the data centres, the chips and the power. That is where the 99's claim must reach, and it is where the next chapter begins.

The tests

This chapter tests the first element of the design, one borderless society, against the owners' exit; the fourth element, every person a shareholder, against the possibility that the owners could leave with the machines; and the floor, against the danger of a herd that is fed but not needed. It also tests the design's own commitment to exit, the fourth test of this book, against the owners' use of it.

Precedent. Elites that withdrew from their obligations while keeping their claims — the French nobility, the Irish absentee landlords, the rulers of rentier states — weakened the bonds that made them tolerable and remained dependent on those they had abandoned. The withdrawal usually ended badly for them, but often only after great suffering for everyone else. Rentier states show that a government that does not need its people's work can afford to ignore them, unless democratic institutions and broad ownership of the wealth came first.

Evidence. The doors are real and in use: private jurisdictions such as Próspera, purchased citizenships, tax exits and the movement of the very rich between countries. Forecasts that AI will replace most human work are made by some of the industry's leaders but remain forecasts; under this book's assumptions they are wrong, which makes acting on them dangerous rather than impossible.

Capture. Private cities are the most direct form of capture in this book: the owners take not the state's institutions but a piece of its territory, write their own law, and use international investment treaties to punish the state that tries to take it back. The trade in passports turns citizenship itself into a commodity.

Exit. This chapter shows that exit is not one thing. The right of a person to leave a society that oppresses them is one of the fundamental freedoms this book defends. The ability of the owners to take with them the wealth that was produced in common, and to escape the obligations that come with it, is something else: it is the farm's oldest escape from the many.

The ledger

What survived. The first element of the design, one borderless society, gains strong support. Every door described in this chapter depends on competition between jurisdictions: havens for wealth, passports for sale, territories for private rule. A world in which those doors were closed — by coordination between states, or by a single society — would remove the owners' newest escape. The floor survives, and becomes more urgent: if any significant part of the population could become unneeded by the owners of the machines, a floor that depends on the owners' goodwill would be the most fragile thing in the design.

What had to change. The design must distinguish two kinds of exit. People must keep the right to leave: to move, to emigrate, to opt out of the society's shared story, as Chapter 7 insisted. But no one may leave with the commons. Wealth produced with public machines and public knowledge must stay in common ownership whatever happens to the people who managed it, and that includes the physical layer of data centres, chips and power on which AI runs; citizenship and residence must not be for sale; and no private party may acquire the right to govern territory or people under rules of its own making. The design must also ensure that the 99 remain needed — or, if the wall proves weaker than this book assumes, that their claim on the common wealth does not depend on being needed at all. That is the strongest argument in this book so far for the fourth element: if every person owns a share of the machines, the owners cannot leave with them, and no one becomes merely a cost.

What remains unknown. Whether the wall holds — and therefore whether the owners' fourth door is an illusion or a real way out. Whether states will cooperate to close the other doors or compete to open them. And whether a society with no outside, which would close every door, could also keep the right of individuals to leave; Chapter 12 will have to answer that.

Bridge: if the machines belonged to everyone

The owners' exit is dangerous because the owners can take the machines with them — the capital, the models, the data, the wealth they produce. The obvious answer is that the machines should not be theirs to take. That is the first element of the design's Part IV, and the oldest and most dangerous idea in the history of the farm: that the means of production should belong to everyone. It has been tried, and the first book recorded what happened when it was tried in the Soviet Union: the owners were removed and the managers became the new owners. The next chapter begins the design itself. It asks how the machines and the models could be owned in common without creating a new class to own them in everyone's name — setting the Soviet lesson against the cases, from Norway's oil fund to the cooperatives of Mondragon, where common ownership has stayed broad.