Iqbal Latif
Third, and this is the heart of it, the world loses its customers. An economy runs on demand, on people who earn and then spend. AI is already reducing the need to pay human beings for thinking work. I have seen it in my own business: legal letters and filings that once needed senior lawyers are now drafted with a machine. Lawyers and consultants will need fewer hands. There will still be a great demand for people who serve, who cook, clean, carry, care and drive, until the robots arrive for those jobs too. As wages shrink, one of the great tax bases of the modern state shrinks with them, and taxes must move to consumption, to capital, or to the owners of the machines. Fewer wages mean less to spend, unless something else replaces them, and the something else is the check that nobody has shown how to pay for. And a robot, however clever, buys nothing. It has no wants, no family, no hunger and no pleasure. The more robots do the work, the fewer people have money to buy what the robots make. Henry Ford’s factories needed buyers, and higher wages helped create them. Musk is building the factory and removing the buyer.
Fourth, the cradle empties faster. He tells us we need robots because the birth rate is falling. But a young couple has children when they feel secure: a job, an income, a future they can see. Take those away and fewer children are born. Put his robots beside the long fall in human numbers, and the two curves meet in a strange place: more machines every year, fewer people every year, until each human being is waited on by several robots and there is almost nobody left to wait on.
So the world divides into two economies. The first is small: the people who own the machines, the models, the rockets, the satellites and the rights. The richest 1% already hold close to half the world’s wealth, and this future hands them the rest of the means of roduction, including the wires through which the rest of us speak. The second is everybody else, living at the edge of the machine economy, not needed to make anything and paid a check to consume what it makes. And what happens to their savings? Their pensions are invested, automatically, in the shares of the machine economy that no longer needs them. The periphery pays for the centre. The many fund the few, and the few call it abundance.
His answer to all of this is that money will not matter. Grant that too. Money is only a claim on real things. Robots can make the claim free. They cannot make the steel free, or the energy, the chips, the launch pads, the factories or the years. I have written before that real wealth grows only as fast as concrete can be poured and steel can be laid, while paper wealth moves at the speed of a mood. Take money away, and scarcity does not vanish. Someone still has to decide what the robots, the mines, the power stations and the rockets are used for: Mars or Manchester, orbital computers or a grid for Africa, a robot for every home or a home for every family. If money no longer makes that decision, something else must. History offers only two answers at this scale. The first is total war. By 1944 American military spending took roughly two fifths of everything the country produced. It lasted four years, and it was paid for with rationing, conscription, war bonds and a nation told to sacrifice. The second is the planning ministry, where the state decides what is built and who works. At the height of Apollo, the greatest peaceful effort of the last century, the whole of NASA took about 4.4% of the federal budget, and it was cut soon after the flag was planted on the Moon.
Follow it to the end, and there are only three ways it can finish. The first is that the arithmetic fails, as I believe it will, and the promise ends the way NEOM is ending. Then the loss does not fall on the man who sold the vision. He has already been paid, in control and in shares. It falls on a hundred million savers: a little on each of them, never enough for any one of them to fight. The gain was concentrated and the loss is spread thin, and because it is spread thin, the next story can be sold again. The second is that it half succeeds. Robots do most of the work, satellites take a large share of the world’s communications, and a handful of owners hold the machines. Then a small class owns everything, a large class lives on checks and has no say in what is built, and the Treasury prints the checks at the pleasure of the owners. That arrangement had a name in the Middle Ages: feudalism. The man who says money will not matter holds 85.1% of the vote in the company that builds the rockets and the satellites, and his shares vest on market value, not on generosity. The third is that the state steps in to decide who gets the steel, the power and the robots, because money no longer can. That is the war economy, run permanently.
Not one of these three is the paradise on the poster. A man who built his fortune in the freest capital market on earth is describing a future that could only be paid for by a war economy or a planning ministry. The kind of economy his future needs is the kind of economy he would never agree to live in.
And there is one more fact, which may be the most important sentence in this essay. In its own accounts, SpaceX judged both sets of milestones in his share grants, the Mars colony and the 100 terawatts in space, improbable, and recorded no expense for either. The company that sells this future to pension funds prices it at nothing in its own books.
The arithmetic tells us the bill. It does not tell us the loss. I spend hours every day working with a machine. It checks my figures, finds my sources and argues with me until the thought is clear. It is the best tool I have ever used. But it is a tool. If I spent those same hours with it for company, for comfort or for love, people would rightly say I had lost my mind. So why are we being sold a future in which that madness becomes normal?
Today, while I was out walking, my wife called me. In the building where I have lived for 20 years, 140 flats drain through one pipe, and something had blocked it. Thousands of gallons of waste were backing up into the basement. I went, I saw it with my own eyes, and I stayed until it was solved. A machine could perhaps have cleared the pipe. But no one would have called a machine, and no machine would have cared whether 140 families went to bed with clean water. Not long ago, by the Thames at high tide, I saw a small child standing at the very edge of the bank. I went to his parents and asked them to hold him. That is what a human society is: people who answer the call for one another.
I still stop to admire beauty when it passes: a helicopter landing, a river in flood, a child kicking a football, a stranger walking by. A machine can compose a tune, but it cannot feel the tune. It can cook a dish, but it cannot taste it or share it at a table. Artificial music and artificial food made by machines, for people who no longer need each other, is not a civilisation. The machines will not take our place by force. I have never believed that, and I do not believe it now. The danger is that we hand our place to them, one comfort at a time: served, fed, entertained and told to sit quietly, until the touch of another person has nothing left to do. We are not born to be served. We are born to create, to struggle, to work and to love. Ten billion robots serving a shrinking humanity is not an economy. It is a waiting room.
I will change my mind the day his company publishes five numbers. The cost of a useful computation in orbit. The cost of a tonne brought back from the Moon. The cost of keeping one person alive on Mars for one year. The cost of carrying one gigabyte to a phone in a crowded city from orbit, against the tower on the street. And the funding plan: whose savings pay, for how many years, and what the savers get back if money no longer matters. Bring the arithmetic, and my retraction will print in the same size type as this charge.
I granted him everything. I granted him the rockets, the robots, the satellites, the lenders and the credit. And the arithmetic still does not close. Not because the rockets fail, but because the bill arrives before the paradise does, and it is sent to us.
I wish this is wrong.
Th Balance Sheet of the Earth, 2027 to 2050
Every company that asks for money must show a balance sheet. Nobody has shown one for the planet that is being asked to pay for this future. So here is mine, in plain words, for the 23 years from 2027 to 2050, the year he has named for a city of 1 million on Mars. Every figure comes from the essay above.
What we have. The world earns about $124 trillion a year. Over the next 23 years, growing at a normal pace, it will earn about $4,000 trillion in total. Of that, about a quarter, $1,000 trillion, at today’s share of investment, is what the whole world spends on building anything: houses, roads, power, hospitals, schools, factories, and replacing what wears out. That is the building budget of the planet for a generation. Behind it stand about 8.2 billion people, 3.5 billion of them working, whose wages pay the taxes, fill the pensions and buy the goods. Its pension funds hold nearly $60 trillion. It makes about 30,000 terawatt hours of electricity a year.
What his fantasy demands. Ten billion robots by 2040, built at about 700 million a year, and then about 700 million a year to replace them as they wear out, so that by 2050 the world has built some 19 to 20 billion in all: $400 to $500 trillion. Computers in orbit big enough to use five times all the energy humanity has today: $100 to $200 trillion just to launch. A city of 1 million people on Mars, a mine on the Moon and a million satellites replaced every five years: several trillion more, plus a fuel factory on Mars and a Martian city that nobody has priced. Total: about $500 to $700 trillion. And on top of that a check for every person alive, which at $30,000 a year each, my figure and not his, would cost $246 trillion a year, twice what the whole world earns.
The gap. On these assumptions, his future takes half to two thirds of the world’s projected investment at today’s share, for 23 years. That means half the houses, grids, schools and water plants that the world would otherwise build do not get built. The check for everyone cannot be paid at all: from 2036, the year he says money stops mattering, to 2050, it would cost about $3,700 trillion, more than everything the planet will earn in that time, and the wages that would have paid for it are the wages the robots take away. His robots alone would use almost all the electricity the world makes today, before the orbital computers are switched on.
What we get back. A settlement of 1 million people on Mars. Industry on the Moon. Vast computing in orbit. Ten billion humanoid machines. Each may have value. What has never been shown is the return on the capital needed to build them: what the Moon sells back to Earth, what a calculation in orbit costs against one on the ground, what a Martian city makes that pays for its imports, and who buys the output of 10 billion robots when human wages are gone. And the one business with a market we can measure, the satellites, could take every phone bill on earth, about $1 trillion a year, and still bring in barely a twentieth of the $20 to $28 trillion spent every year to reach it.
What the same money buys here. Power lines from one end of the earth to the other, carrying sunshine from the deserts, for $2 to $3 trillion. Fresh water for every dry coast in Africa and Asia for a few trillion more. Every school, clinic and grid the poor world lacks, and the replacement of every worn out thing the rich world already has. All of it creates things that people alive this century will use: electricity, water, schools, clinics, homes and work, on the planet where 8 billion people already live.
Who signs. A banker who saw this on his desk would ask one question before anything else. The answer is that nobody signs. The pension funds bought the shares because the index funds they hold were required to, the Treasury prints the checks because a speech told it to, and the man who drew up the plan holds 85.1% of the vote, while the company itself recorded no expense for his milestones because it judged them improbable.
In one sentence: on these assumptions, he asks the world to spend half of its projected investment for a generation on things whose return has never been shown, paid for by people whose jobs the plan removes, and nobody has been asked to sign.
