The most underrated story in geopolitics is that almost the entire world is about to get old at once, and almost nobody has a plan for it. Japan is the case everyone points to, because Japan has been running the experiment since about 1990 and the results are in. You can have a peaceful, rich, technologically advanced society that is shrinking and graying, and it can hold together for a long time, and it will feel like a country in late autumn. Not much grows, and not much moves. The young spend their working lives holding up a very large old generation, they put off families of their own or skip them, and the next turn of the wheel comes in smaller still.
What is new in 2026 is that Japan is not the outlier anymore. South Korea's fertility rate is 0.72, which is not a typo and is the lowest figure ever recorded in peacetime, low enough that each generation arrives at roughly a third the size of the one before. China's population peaked in 2022 and is now falling. Italy and Spain have sat below replacement for thirty years. Germany keeps its numbers up through immigration, and the politics around that immigration have turned ugly in the way everyone could have predicted. Even India, the great exception, the country that was supposed to stay young and growing for decades, slipped below replacement fertility around 2020. We are just starting later, and from a much poorer base, which is its own problem: India is on track to grow old before it grows rich, and Japan at least did these things in the right order.
Almost all the economic machinery we use to think about the future quietly assumes the population keeps growing. Pension systems, housing markets, equity markets, the whole idea of growth as the point of policy, all of it rests on the pyramid staying a pyramid, with enough young workers arriving at the base to hold everyone above them up. Invert the pyramid and the arithmetic stops working, and every choice left on the table is grim. Cut benefits to the old, who are many and who vote. Raise taxes on the young, who are few and who work. Import workers, which fixes the economics and inflames the politics. Or borrow the gap, which is what most governments are quietly doing until the day they can't.
The cultural side of this matters as much as the economic side and gets far less attention. A society with a lot of young people feels different from one with very few. The baseline energy is different, the tolerance for disruption is different, the appetite for starting something, a business, a movement, a war, a family, is different. Older societies are more cautious, more nostalgic, more protective of what already exists, and more inclined to vote down anything that costs money or changes the view from the window. You get cities that feel like museums and governments that feel like hospice care. The hill towns of southern Italy, where the median age is past fifty and the grandchildren have long since left for Milan or Munich, have a specific quality of quiet to them. It is the quiet of a place that has already had its future.
Here is the fork in the road, and it is labor productivity. Demographic inversions used to be rare and catastrophic, and the closest precedent we have is the Black Death, which killed something like a third of Europe and then raised real wages for the survivors for more than a century, because labor had suddenly become scarce and capital had to bid for it. Serfdom fell apart in large part because landlords no longer had the leverage to hold people in place. Scarce workers meant power flowed to workers. That is the conventional model of what a shrinking workforce does: wages rise, conditions improve, capital gives ground.
Everything about this inversion breaks that model, because the productivity gain that should have flowed to scarce workers is being captured instead by AI systems that draw no wage. Fewer workers multiplied by far higher output per worker can produce two completely different societies, and which one you get depends entirely on who owns the multiplier. In one, the work week shrinks, prosperity spreads, and an aging population is treated with some grace because labor is still valued. In the other, a smaller number of workers do more for the same pay while the surplus flows to whoever owns the model weights and the robots, and the displaced older population waits on a pension that gets quietly rewritten every few years. We are on track for the second one by default, because the ownership of AI is highly concentrated and the political coalition that might redirect the surplus does not exist yet.
This is the trade sitting underneath the whole aging debate. Japan's version stays relatively gentle because Japan still pushes its productivity gains back out through institutions, lifetime employment, a compressed income structure, executives who are not paid four hundred times their workers. The gains mostly stay with the people who produced them. The American and British versions of the same demographic problem will go differently, because there the surplus gets captured by whoever owns the automation, and neither political system has any real mechanism for sharing it back at the scale a top-heavy population needs. You end up with a small class who own the productive capital, a shrinking and anxious middle that knows it is one model update from redundancy, and a large older population leaning on transfers the working population cannot fund and does not want to.
The conversation that might head this off is not seriously happening anywhere. No major party in any rich country has a coherent story for how the gains from AI get shared out. Where the debate exists at all, it is still stuck on whether AI will take jobs in the first place, which is an argument about the premise. The actual question, what happens to the surplus, is barely being asked, and by the time the answer is impossible to ignore the ownership will already be locked in. Rewriting who owns what after the fact tends to be either impossible or violent, because that is how these things have always gone.
The countries that come through the next thirty years in the best shape will be the ones that solve both halves at once, a functioning immigration system to soften the demographic slope and a real political settlement over the AI surplus so it does not all pool at the top. The list of places that look capable of managing both is short. Most of the rich world is going to muddle through instead, with worse politics, more debt, and shrinking ambition, and it will blame the result on almost anything other than the two things actually causing it.