For almost everyone, a job is how you earn the money to live. Robots and AI are about to do most of the jobs — and take those paychecks away. The usual answer is a universal basic income. The usual objection is that we cannot afford it. This page argues that both sides are looking at the wrong number — that the way to make UBI affordable is not to raise incomes but to collapse the cost of physical goods, and that the way to do that is to make almost everything inside the city that uses it.
The chain of the argument: automation removes labor → a UBI is needed to replace wages → but a UBI only works if goods are cheap → goods get cheap when you delete factory labor and long-haul freight → you delete freight by building in every city → which shrinks global shipping to raw materials, energy, and luxury → which turns the world economy from a trade-in-goods system into a trade-in-resources system. This document tries to put rough numbers on each link.
Apart from the sourced UBI explainer near the top, every figure on this page was estimated by Claude, an AI assistant, from public rules of thumb and back-of-envelope arithmetic. None of it has been reviewed by an economist, and much of it is simplified to the point of being wrong in the details. Real economies are dense with feedback loops, prices, and politics that a page of round numbers cannot capture. Ranges are wide on purpose.
This is a thought experiment meant to make one specific idea concrete enough to argue with — not a forecast, not policy, and not investment advice. The value is in the structure of the argument, not the third digit of any number. If a figure here matters to you, check it yourself.
No economics needed. If the charts and big numbers below look like a headache, read this part and you’ll understand the entire page.
Machines are about to do most of the jobs. Not just factory work — computers and robots are getting good enough to drive, deliver packages, answer phones, do paperwork, and handle a huge amount of office work too. Over the next couple of decades, a lot of the jobs people do today will be done by machines instead.
That creates one big problem. The way a normal person gets money is by having a job: you work, you get a paycheck, you use it to pay for food, rent, and everything else. No job means no paycheck, and no paycheck means no way to live. So if machines take the jobs, most people are left with nothing — even though the machines are actually making more stuff than ever before.
One idea to fix this is called UBI. It stands for “universal basic income,” and it just means giving every single person a set amount of money every month — say a thousand dollars — simply for being alive, whether they have a job or not. Think of it as a monthly allowance for everybody.
The usual complaint about UBI is that we can’t afford it. Giving every adult a thousand dollars a month adds up to trillions of dollars a year. Where would all that money come from? Normally, that’s a fair objection.
Here is the twist this page is about. Instead of only asking “how do we give people more money?”, it asks “how do we make everything so cheap that people don’t need as much money?” If a phone, a chair, or a shirt costs a small fraction of what it does today, then even a little money goes a very long way.
So why is stuff expensive in the first place? A big part of what you pay isn’t the thing itself. It’s all the workers who had to be paid to make it, the cost of shipping it halfway around the world on giant boats, and all the stores and middlemen who each take a cut. Remove those and the price mostly falls apart.
How do you remove them? Robots, and making things locally. If robots do the making, you’re not paying a whole factory of workers. And if the thing is built right in your own city instead of shipped from another country, you’re not paying for the long journey either. What’s left is basically the raw materials, the electricity, and the machine — so the price could drop by two-thirds or more.
For that, every city needs its own little robot factory. A local place that can make most everyday things close to home, instead of one giant factory in a faraway country. The design for a product (basically the recipe) gets sent over the internet instantly; the actual object gets built locally. The page estimates the United States would need around 1,400 of these factories, and the whole world about 32,000.
It even shrinks down to your garage. The same machines get cheap enough that regular people can keep one at home or in a small shop — like a super-powered 3D printer — and make things for themselves and their neighbors. Almost anyone could become a maker, or start a small company, without a factory or a lot of money.
Because cities make their own stuff, the giant cargo ships mostly stop. We’d no longer ship finished products across oceans — only raw materials (like metal and oil) and a few fancy luxury items. That’s also good for the planet, because those ships burn a lot of dirty fuel.
Now the money math works. If everyday things are really cheap, a basic-income check buys far more than it would today. And the money to pay for that income comes from the robot factories themselves — they make so much, so cheaply, that a share of it can be handed to everyone.
But here is the huge catch: who owns the robot factories? If a few rich companies own them, those companies keep all the money and everyone else is still broke — that’s the bad ending. If the factories are owned by regular people — like a community business, or a club that every citizen is automatically a member of — then everyone shares in what the machines make — that’s the good ending. This single question decides whether this future is wonderful or awful.
And no, this isn’t just “socialism.” Socialism is the government owning and running everything. Here, regular businesses and regular people still own things, buy, sell, and compete like always. There are only two differences: a money floor so nobody falls into poverty, and the big factories are owned by the citizens themselves, not by the government or by a handful of billionaires.
What life might feel like. With cheap goods and a basic income, a family might not need both parents working two jobs anymore. People would have more free time — for their kids, hobbies, working out, music, art, sports, and just being around their neighborhood. Work becomes something you do because you want to, not because you’d starve otherwise.
One last, honest thing. The writer isn’t claiming to know the future. Every number in the thought experiment is a rough guess made to explain an idea, and plenty of it could turn out wrong. It’s here to make you think, not to be treated as fact. Everything below is simply this same story, told in more detail and with the numbers shown.
Real programs and trials, checked against their published results in October 2026. Unlike the thought experiment below, these are not Claude’s estimates.
Universal basic income (UBI) is a regular cash payment to every adult, with no work requirement and no means test. No country pays one big enough to live on today. What exists are smaller payments and time-limited trials, and most trials pay some people rather than everyone, so researchers often call them guaranteed income. Here is what four of them found.
Nearly every Alaska resident gets a yearly share of the state’s oil-funded savings fund: $1,000 in 2025, paid to more than 600,000 people. It is the closest thing in the US to a universal payment, but it comes once a year and is far too small to live on.
2,000 unemployed people aged 25 to 58, picked at random, got €560 a month for two years with no strings. Work barely moved: about six more days of employment than the comparison group from November 2017 to October 2018, and even that is hard to separate from a rule change in 2018. Recipients were more satisfied with their lives and reported less mental strain, depression, sadness and loneliness.
125 people in lower-income neighborhoods got $500 a month for two years. In the first year, full-time work among recipients rose from 28% to 40%, against 32% to 37% in the comparison group.
1,000 low-income people got $1,000 a month for three years; 2,000 others got $50 a month for comparison. Recipients worked 1 to 2 fewer hours a week, their partners about as much less, and they were 4.2 points less likely to be working or looking for work. For every dollar received, the household’s other income fell about 28 cents. The biggest gain in time went to leisure.
That bill is the problem the rest of this page plays with: D‑6 asks whether cheaper, locally made goods could make the same payment go much further.
Sources: Alaska Department of Revenue (2025 dividend) · Kela, Finland’s basic income experiment · VATT, employment effects · Results for America, Stockton’s SEED · Vivalt et al., NBER Working Paper 32719 (OpenResearch) · US Census Bureau, 2020 adult population
For two centuries, a wage did two jobs at once. It paid people, and it was the mechanism by which ordinary people claimed a slice of everything the economy produced. You worked, you got money, the money was your coupon to redeem output. When machines do the work, the output still exists — arguably more of it — but the coupon stops printing.
A universal basic income is simply a new coupon: instead of earning a claim on output by working, every person is issued one by virtue of being a person. The standard rebuttal is arithmetic. A modest United States UBI — call it $1,000 a month to every adult — is roughly $3 trillion a year, on the order of a ninth of the entire economy. Fund it by taxing the wages that automation just deleted and the money is not there.
That rebuttal is correct at today’s prices. The move this page makes is to treat prices as the variable, not the constant. If the things a UBI has to buy get dramatically cheaper, the size of the coupon needed to live well shrinks with them — and the surplus thrown off by near-free automated production becomes the thing that funds the coupon in the first place.
The hub is not a car plant that makes one thing. It is a floor of general-purpose robotic cells and printers that switch products in software — furniture on Monday, appliances on Tuesday — so one building covers most of what a household buys.
What crosses borders is the design, weightless and instant. The heavy, expensive thing — the finished product — is made where it is used. This is the single move that deletes freight.
The cleanest version feeds on local recycled feedstock and only imports the raw commodities a region genuinely lacks. Waste becomes the primary mine; shipping is reserved for what the ground nearby cannot provide.
Delete labor and freight and the remaining cost is materials plus energy. The whole idea only works in a world of abundant, cheap, clean electricity. If power is scarce, none of this works.
The city hub is the industrial tier, but the same crash in the price of automation does not stop at the city. The tools that let a hub run without workers — capable robot arms, desktop machines that print in polymer and metal, and an AI that turns “I need a replacement bracket for this” into a finished machine file — get cheap enough to sit in a garage, a spare room, or a small leased warehouse bay. The average person becomes, quietly, a small manufacturer: making for themselves, their family, and the few blocks around them from the same open catalog of designs the hubs use.
This is the long tail of localization. The city hub handles what genuinely needs scale, heat, or hazardous processing; the garage and the corner workshop handle everything bespoke, urgent, or personal — the one-off part, the custom furniture, the small run a neighborhood wants. Production stops being something that happens in distant countries and becomes something that happens on your street, and sometimes in your driveway. It is the return of the local craftsman, except the craftsman now commands a small robotic crew.
It also detonates the cost of starting something. Today, turning an idea into a physical product means tooling, minimum order quantities, an overseas factory, and capital most people never have. When the design is an AI conversation and the manufacturing is a machine in the corner, the distance from “idea” to “a product in a customer’s hands” collapses to days. Expect a surge of one- and two-person product companies — people designing, making, and selling their own goods, from custom furniture to specialized tools to small-batch electronics, with no factory, no importer, and no warehouse behind them. Entrepreneurship stops being reserved for the well-capitalized; being a maker of things, not just a buyer of them, becomes something almost anyone can do.
At roughly a billion dollars a hub, localizing the United States is on the order of $1.4 trillion, and localizing the planet is on the order of $32 trillion. Those sound impossible until you divide by time. Spread the American build over twenty years and it is about $70 billion a year — roughly a quarter of one percent of the economy, smaller than what the country already spends subsidizing things it cares less about. The global build over the same span is around $1.6 trillion a year, roughly one and a half percent of world output and a few percent of what the world already invests in fixed capital annually.
The binding constraint is not money and not steel — the concrete and steel for thirty thousand hubs is a fraction of a single year of world production, spread across two decades. The constraint is robots, chips, motors, and batteries, and the copper, lithium, and rare earths behind them. To build and run this many automated hubs, the world’s output of industrial robots and power electronics has to rise by something like two orders of magnitude. That, not the buildings, is the real twenty-year project — and it is the same bottleneck as the clean-energy transition, which is convenient, because this plan is completely dependent on that transition succeeding.
Both halves of this — the hubs to build and the people to support — scale with population, so the burden falls hardest on the most populous nations. India, China, and the countries of Africa each need thousands of hubs and must stand up a floor for well over a billion people apiece. The rich, smaller countries have the easy arithmetic; the giants carry a load an order of magnitude larger, on timelines that are political as much as industrial. Population, in this model, is the thing that sets how heavy the lift is — and the heaviest lifts sit with the countries least able to coordinate them quickly.
And a handful of countries do not merely carry a heavy load — they lose their entire economic model. Nations like China built their modern rise on making goods for foreigners: an enormous share of output is factories producing objects that cross an ocean to be sold in someone else’s country. In a fully localized world that demand does not shrink, it goes to zero — every country makes its own goods, so no one needs a distant workshop to make theirs. The export-manufacturing economy is not reduced or reshaped; it is deleted. The countries that leaned hardest on being the world’s factory face the sharpest pivot of anyone: away from selling labor sealed inside objects, and toward selling the only things that still travel — raw resources, energy, designs, and expertise.
Today a country gets rich mostly by adding labor to other people’s materials — importing components, assembling them cheaply, exporting the finished good. Delete the labor step and that entire business model evaporates. What is left with any leverage is being upstream: owning the materials and the energy. In a localized world the valuable exports are ore, lithium, rare earths, food, and electricity — not televisions.
This is the part that should worry you as much as it excites you. A resource economy concentrates power in whoever sits on the resources. It could free import-dependent nations from the factories of others, or it could hand extraordinary leverage to a handful of resource-rich states and the firms that own the mines. Who owns the hubs and the mines decides whether this ends in shared abundance or in the most lopsided economy in history. The technology is neutral; the ownership is everything.
Because every country’s GDP has a different shape, the same change lands unevenly. The United States is trade- and service-heavy, so it loses large chunks of GDP in retail, finance, insurance, and logistics, and its white-collar work is hit hardest by the AI half of the shift; but it holds real cards — energy, agriculture, robotics and AI intellectual property, and a big domestic market to localize into.
The clearest losers are the export-manufacturing economies. China most of all — making goods for foreigners is on the order of a fifth of its output, and in a localized world that demand goes to zero — with Germany, Japan, South Korea, Vietnam, and Mexico close behind. The trade and shipping hubs (Singapore, the Netherlands, Panama) lose the middleman role their economies are built on.
The clearest winners are the resource states. The Gulf, Australia, Chile and the DRC (copper, cobalt, lithium), Canada, Brazil, and Russia gain leverage, because energy and raw materials become the one thing every local hub still has to buy from somewhere. And the experience economies — tourism, hospitality, and culture — quietly rise everywhere, because the thing this whole system hands people is the two ingredients leisure spending needs: time, and a floor of income.
The moment a reader sees a universal income funded by a public dividend, one word comes up: socialism. It is worth taking seriously, because the resemblance is real on the surface and the differences underneath are enormous. The short version: socialism is a choice about how to run an economy where humans do the work; this is a forced adaptation to an economy where they mostly don’t. One is an ideology adopted for moral reasons; the other is simply what a market economy has to do when its own distribution mechanism — the wage — stops functioning.
None of this makes the two unrelated. Both put a universal floor under everyone, both rely on some collective claim on production to pay for it, and both live or die on the ownership question this page keeps returning to — push the “public stake” far enough and you do arrive somewhere that looks socialist. So the honest position is that the label is a political fight, not a technical one. You can reach the identical conclusion — that a basic income becomes necessary — from the free-market right (the market’s wage mechanism broke, so patch it and keep the markets) or from the redistributive left (at last, shared abundance). What matters is not which flag it flies under but whether the mechanism actually works: cheap goods, a funded floor, and markets still doing the allocating.
If ownership is the whole game, the obvious question is: owned by whom, exactly? The failure mode is concentration — a few corporations, or a single government, ending up holding the hubs. The good news is that there are already proven ways to own productive infrastructure that are neither corporate monopoly nor state control. Here are the ones that could build and run the hubs.
A city or county builds and operates the hub as a public utility, the way it already runs water, transit, and libraries. Residents are the beneficiaries, the surplus returns as a local dividend or as services, and the asset is accountable to local voters and cannot leave town.
The hub is owned by the people it serves — one member, one share, one vote — with profits returned to members rather than to distant shareholders. This is the co-op and union model, and it is not theoretical: America’s rural electric cooperatives, its credit unions, and Spain’s Mondragón federation already run at scale this exact way.
Individuals and small firms build and run hubs as ordinary businesses, competing on price, quality, and speed — right down to the garage and neighborhood-workshop tier. Plain market competition, now open to almost anyone because the cost of the tools collapsed.
Most likely none of these wins outright. Municipal hubs, member-owned co-ops, and private operators coexist and compete in the same city — the way public, charter, and private schools already do. Competition between ownership models is itself a check on any one of them running away with it.
All of that still leaves the door open to the failure mode this page keeps warning about — a few corporations quietly buying up the hubs, or a government nationalizing them. The structural fix is to take that door off its hinges: a national rule that the hubs must be owned by the citizens directly — not by the state, and not by concentrated private capital. Think of it as a union that every citizen is automatically a member of.
In practice that looks like a citizens’ trust: the productive hubs are held in common on behalf of every citizen, each person holding a single non-transferable membership share that cannot be bought, sold, or piled up. Governance is democratic — members elect the boards that run the trust — and the surplus is paid out as the basic-income dividend. Crucially, it is owned by the people, not the government: politicians administer it no more than they administer a credit union or a rural co-op. This is the line that separates the whole idea from socialism on one side and oligarchy on the other — the machines are collectively owned, but by the citizens as members, never by the state and never by a handful of shareholders. It is the difference between a commons and a command economy — and between a commons and a monopoly.
The dual-income household was, in part, a response to rising prices for the essentials of a decent life: housing, childcare, education, healthcare. Manufactured goods getting cheaper does not fix any of those — but a UBI plus cheap goods changes the math of the second job. If a household’s basic goods cost a fraction of what they do now, and a basic income covers the floor, the marginal reason for both adults to sell forty hours a week weakens.
The likely early pattern is a quiet reversion: many families drop back to one earner, with the other doing the care, community, and creative work the paid economy never priced well. Further out, as automation deepens and the dividend grows, paid work itself becomes optional rather than obligatory — not because it is banned, but because survival no longer requires it. That is either the best thing that ever happened to families or a crisis of meaning, depending on whether people find purpose outside a paycheck. This page cannot tell you which.
Take away the obligation to sell forty or fifty hours a week and the real question is what fills them. The likeliest answer is also the most human one: the things people already reach for the moment work lets go. More time in the gym and on the trail. More hours in games, alone and together. More music made and not just streamed, more painting, more writing, more poetry — the arts that a wage economy priced at roughly nothing suddenly have the one thing they always lacked, which is time.
And more of the unpaid work that actually holds a society together: coaching a kid’s team, cooking for people, building things for the joy of building them, sitting with aging parents, knowing your neighbors. The care and creative work the market never valued becomes the main event rather than what people squeeze into the exhausted edges of a workweek. There is a real risk here — jobs bundled income together with meaning, structure, and status, and not everyone flourishes when that bundle comes apart. But the direction of travel is a civilization that spends its abundance on being human instead of on being employed.
And there is a quieter change underneath all of it: people are simply around more. Home in their neighborhoods during the day instead of commuting out of them, present on the streets that used to empty at 8am. The corner shop, the park, the local team, the block that actually knows itself — the “third places” and the everyday neighborliness that long work hours hollowed out have the room to grow back. A society where most people are present where they live, rather than absent at jobs across town, becomes a more connected one almost by accident — a return of community that no policy could order but that free time quietly produces.
One localized manufacturing hub per ~250,000 people puts the United States at roughly a thousand to a few thousand facilities — fewer than it has big-box stores. The planet is roughly thirty thousand.
~$1.4T for the US and ~$32T worldwide, which spread over two decades is ~$70B/yr and ~$1.6T/yr — fractions of a percent to a couple percent of output. The real bottleneck is robots and chips, not money.
Materials and energy remain; the wages, the ocean voyage, and the middlemen do not. Mass-produced goods fall toward a quarter of today’s price — not zero, but enough to change what a basic income can buy.
Finished-goods shipping largely ends; bulk ore, grain, and energy keep moving, plus a thin luxury flow. Global trade re-forms around raw materials and electricity — a resource economy.
A fixed UBI buys roughly three times as much once goods collapse, and it is funded by the surplus of the automated hubs rather than by taxing vanished wages — if, and only if, that surplus is publicly captured.
Cheap goods do nothing for the cost of a place to live, and housing is where a basic income is most easily swallowed. Hand every renter a monthly check and the obvious danger is that landlords simply raise the rent to absorb it — the floor you gave the tenant becomes the new floor of what the landlord can charge, and the person is no better off than before. This is the single most likely way the whole plan quietly fails: not with a crash, but with the UBI vanishing into rent.
Two forces push the other way. First, when income no longer comes from a job in a particular city, people can live anywhere — the pressure on expensive job-center metros eases as work decouples from place, and demand spreads out to cheaper towns. Second, the same automated, localized construction that makes goods cheap can make building cheap: robotic and modular construction attacks the cost of new housing supply head-on. But neither happens on its own. For a UBI to secure shelter instead of inflating rents, it has to arrive alongside real housing policy — building far more homes, taxing land rather than the buildings on it so that hoarding is punished, and keeping some floor of social or public housing outside the speculative market. Handing out income without fixing housing supply just hands the income to whoever owns the housing.
The development banks would have to be rebuilt. The World Bank and the IMF were designed for a growth-through-trade world: lend a poor country money, it builds export industries, it grows and climbs the value chain and repays the debt. A localized, UBI-floor world voids every assumption in that model — there are no export industries to build, and the goal is not growth but a floor. The institution that finances development would have to become something closer to a global dividend office: paying to build hubs in poorer nations as grants rather than loans, accounting for countries in resources and energy instead of GDP, and administering the global resource dividend that funds a basic-income floor where a nation cannot yet fund its own. Debt-for-development gives way to build-the-floor-for-everyone — a different institution wearing the same name.
And in the United States, the billing has to go. A basic income cannot secure health while health is priced the way it is now. Something like a quarter to a third of American health spending is not care at all — it is billing: coding, claims, denials, appeals, and the small armies on both sides who fight over them, on the order of $800 billion to $1 trillion a year of pure administrative friction. Cheap goods do nothing for this, and neither does a check handed to a patient, because the invoice on the other side is written to capture whatever it can. For a UBI to actually buy health rather than be swallowed by it, the insurance-billing layer would have to be removed outright — replaced by direct public provision at a set price — so the floor a person is given is not instantly consumed by a bill engineered to exceed it. Of everything on this page, this may be the change with the fiercest political resistance and the plainest necessity.
Ownership is the whole game. Every number above assumes the surplus from the hubs is shared — publicly owned, taxed, or dividended. If the hubs are privately held and lightly taxed, the same technology produces the most concentrated wealth in history and no UBI at all. Nothing in the physics decides this; politics does.
Energy is the floor. Delete labor and freight and cost becomes materials plus power. The entire idea rides on abundant, cheap, clean electricity. If the energy transition stalls, localized manufacturing is expensive, not cheap, and the price collapse never comes.
Robots and materials are the near-term wall. Scaling automated hubs by a hundredfold means scaling the world’s output of robots, chips, motors, and batteries — and the copper, lithium, and rare earths behind them — against real geological and geopolitical limits. This is the constraint that sets the true timeline.
Scarce things stay scarce. Cheap goods do not make cheap land, cheap housing, cheap healthcare, or cheap status. A world of near-free objects can still be a world of brutal competition for the things that cannot be printed — and a UBI has to be big enough to cover those the old-fashioned way.
The transition can fail even if the destination is good. The gap between jobs vanishing and the floor arriving is where the real danger sits. A destination of shared abundance is no comfort to a household that loses its income a decade before the UBI is funded. Sequencing is not a detail; it is the hardest part.
In one sentence
Automation makes a universal basic income necessary; localized, automated manufacturing — by collapsing the price of goods and the cost of shipping them — is what could make it affordable; and the leftover question, the one that decides whether this is utopia or oligarchy, is who owns the machines.