Chapter Eighteen: The Machine That Cannot Move

Stand at the edge of the port of Shanghai, or Shenzhen, or Ningbo, and try to take in the scale of it. You cannot, really; it exceeds what the eye is built to hold. The container cranes march to the horizon in both directions, each one taller than a fifteen-story building, and beneath them the steel boxes are stacked in a city of their own, a geometric landscape of millions of containers moving in a choreography so vast it can only be managed by computer. China's ports move more cargo than any other nation's on Earth, by an enormous margin — its great harbors together handle more container traffic than all the leading ports of the rest of the world combined. It is, by some distance, the largest, most sophisticated, most productive piece of industrial machinery that human beings have ever assembled.

And now notice the one thing about it that matters most for this chapter: it cannot move.

This is the deep, defining asymmetry between the two things this part of the book has been about. A worker can move. A worker can get on a plane in Manila or Lagos and step off it in another country, fully formed, ready to work, as we saw in the last chapter. But a deepwater port cannot get on a plane. Neither can a national rail network, or a power grid, or a forty-year-old web of ten thousand specialized suppliers clustered within a day's drive of one another. China spent forty years building the most magnificent immovable machine in the history of the world. And it built the entire thing on a single assumption, never stated because it never needed to be: that there would always be hands to run it.

That assumption is now failing. What happens to the greatest machine ever built when the people it requires stop being born is not a sudden collapse but something slower, stranger, and in the end more total.

What China Actually Built

It is tempting to picture China's manufacturing dominance as a matter of factories — buildings full of machines that could, in principle, be rebuilt anywhere. But the factories are the least of it, and the most movable. What China actually built, and what makes it so nearly impossible to replace, is something far harder to see and far more rooted in place.

It built the ports, and the rail and road networks feeding them, and the power to run all of it — physical infrastructure welded to specific coordinates on the map. And it built something even more valuable and even more immovable: the dense, accumulated, decades-deep web of supply-chain knowledge. This is the invisible asset, and it is the real moat. It is the fact that within a single region you can find a supplier for any component, and a supplier for the thing that supplier needs, and the experienced engineers who know how to make the whole network sing — relationships and process expertise and institutional memory built up over forty years and clustered, physically, in one place. When a company tries to move production out of China, it discovers that it can relocate the factory easily enough and then cannot find, in the new country, the thousand surrounding suppliers and the embedded know-how that made the factory work. (Recall the Taiwanese electronics giant — the very company that assembles the world's iPhones — which announced, with fanfare, a vast new plant in Wisconsin, and delivered a small fraction of what it promised: not even the planet's largest contract manufacturer could conjure its supply-chain ecosystem onto new ground. Capital and competence do not relocate on command.)

Economists have a word for this property — embeddedness — the degree to which capital is locked to a location and cannot be picked up and carried elsewhere. And embeddedness is a wonderful thing, a source of enormous advantage, right up until the moment the rationale for the whole enterprise changes. Then it becomes the opposite: a trap. The more deeply embedded the infrastructure, the more completely it is stranded when the conditions that justified it disappear. China built the most embedded industrial machine in history. Which means that when its rationale changes, it will be the most stranded.

The Rationale Is Leaving

And the rationale is leaving, on two legs, both of which are simply the demographic story of this book arriving at the factory gate.

The first leg is that the workers are vanishing. The one-child policy we examined earlier did exactly what such a policy does to an age structure: it produced a generation of young workers dramatically smaller than the generation of older workers it now has to replace. China's working-age population has already passed its peak and is declining; the cohort of twenty- and thirty-somethings who are supposed to run the machine is far smaller than the cohort aging out of it, and there is no larger cohort behind them, because they were never born. The great machine was designed for an endless supply of young hands, and the supply has begun, irreversibly, to contract.

The second leg is that the remaining workers are no longer cheap. This is the demographic dividend running in reverse. For decades China enjoyed the most powerful demographic tailwind in economic history — an enormous young cohort pouring into the workforce, holding wages down and output up, the engine of the entire boom. But a demographic dividend is a one-time windfall, not a permanent condition (we have seen this shape before, in the very first chapters of this book). The young cohort that powered the boom is now the middle-aged cohort heading toward retirement, becoming dependents rather than producers, and as young workers grow scarce their wages rise. The cheap labor that was the whole point is getting expensive at the same moment it is getting rare. The machine that needs young hands is running short of them and being charged more for the ones that remain.

The Decoupling

Now put the immovable machine and the vanishing labor together, and you arrive at the structural heart of the matter, which is a mismatch of an almost cruel geometry. The infrastructure and the available young labor are coming apart — decoupling — and they are decoupling into the wrong places.

The infrastructure is in China and cannot leave. The young labor, increasingly, is not in China and increasingly exists only where the infrastructure is not — in parts of Africa and South Asia that have the youth but not the ports, not the supplier networks, not the forty years of embedded knowledge, and no way to conjure them quickly. So the world is left holding two halves of a machine that cannot be joined: a magnificent industrial apparatus with a dwindling workforce in one place, and an abundant young workforce with no comparable apparatus in another. The infrastructure cannot move to the labor; the labor cannot move, at anything like the required scale, to the infrastructure; and the embedded knowledge that is the real asset cannot be transplanted to a new geography on any timeline shorter than decades — if it can be transplanted at all. This is the supply-side completion of the earlier point about there being no next country. From the labor side, the pipeline has no next link because everyone is aging. From the infrastructure side, there is no next link because the machine took forty years to build and is bolted to the ground.

History has a name for what China's manufacturing base is becoming, and a set of grim precedents for it. A stranded asset is a piece of infrastructure whose economic rationale has collapsed but which cannot be relocated or repurposed — it simply sits, too valuable to abandon and too pointless to run. Lancashire, in the north of England, was once the greatest textile-manufacturing complex on the planet, the very forge of the Industrial Revolution; when production moved to cheaper Asia in the twentieth century, the mills could not follow, and they stood silent and darkening for decades while the towns built around them slowly hollowed out and never fully recovered. The auto plants of Detroit. The steel and coal complex of Germany's Ruhr. In every case the infrastructure outlived its reason for existing, sometimes by half a century, and the human communities welded to it declined not in a sudden crash but in a long, grinding subtraction. China is on course to become the same phenomenon at a scale the world has never witnessed — and for a genuinely new reason. Lancashire was undercut by a cheaper rival. China has no cheaper rival; there is no next country. China is being stranded not by competition but by its own demography. The factories will not be beaten. They will simply run short of the people to run them.

From Exporter to Importer

The most vivid sign of the reversal is what an aging China will increasingly want from the rest of the world. For forty years China's relationship to global labor was that of the great supplier — the place that absorbed the world's manufacturing because it had workers to spare. That is about to invert. An aging China does not need workers to staff export factories. It needs workers to do what every aging society needs done: to care for the enormous and growing population of the old, to staff the hospitals and the care homes, to provide the domestic and personal services an elderly population consumes, and to maintain the vast infrastructure as the population that built it retires. China, in other words, is about to enter the same scramble we saw earlier — the global competition for young, embodied, hard-to-automate caring labor — not as a seller but as a buyer. The country that defined the age of labor surplus is becoming one of the largest symptoms of the age of labor scarcity. The exporter of cheap labor's output is turning into a rival bidder for the scarce labor itself.

And it is not only China's factories that are stranded. Step back and look at the whole planetary apparatus that was built around them. The colossal fleets of container ships; the ports on the receiving end in Los Angeles and Rotterdam and Hamburg; the global logistics firms; the just-in-time supply chains threaded across oceans; the entire intricate machine that the world assembled, over forty years, for the single purpose of moving cheap Chinese goods to wealthy consumers — all of it was built on the assumption of a permanent, enormous flow of cheap manufactured goods across the seas. When the goods stop being cheap, because the labor making them has run out, the machine that exists to move them has progressively less to do. The world trade system is, in the long run, a stranded asset built to serve another stranded asset. We constructed an entire global economy to exploit a labor differential, and the labor differential is disappearing.

The Hope That Closes the Part

There is, of course, an obvious answer to all of this, and it is the answer the entire world is now betting on with a conviction and a sum of money unlike anything in economic history. If the problem is simply that there are no longer enough young hands to run the factories — if the only missing input is human labor — then surely the solution is to stop needing the human labor at all. Automate the factory. Let the machines run the machines. Replace the workers who were never born with robots and artificial intelligence, and at a stroke the stranded infrastructure is rescued, the labor shortage is solved, and the entire demographic crisis that this book has spent eighteen chapters describing is rendered, at last, irrelevant by technology.

This is the great hope. It is the bet underneath the trillions of dollars now pouring into artificial intelligence and robotics; it is the reason the most valuable companies on Earth are the ones promising to build the machines that will think and the machines that will move. The wager is that technology will fill the human gap that demographics is opening — that we can have the productive economy without the people, the output without the workers, the civilization without the next generation. It is the last and largest hope in this book, and almost everyone now alive is, in one way or another, counting on it.

And it is wrong. Not wrong in the sense that AI and robots will fail to arrive — they are arriving. Wrong in the sense that almost everyone has the direction of their effect exactly backwards. Artificial intelligence is not the rescue from a shrinking population. It is the force that turns a shrinking population from a slow problem into a fast one, because — and this is the proposition the next several chapters will make — AI does not create value. It destroys it. It does not generate. It deflates.

That is where we turn next.