What is wealth, really?

Assumption 1: money is wealth

Suppose only two people are left on Earth. You have $10bn; the other person has only clean water, enough food, and a solar panel. Who is richer? The answer is obvious—your $10bn is worth something for one reason only: someone else is still willing to trade real things for it. The moment that person leaves, the cash is just paper.

Money, taken apart, is only an agreed medium for keeping accounts. It cannot be eaten, drunk, or burned for power. Gold, dollars, yuan—all of it runs on belief. So wealth is not money; it is the ability to meet needs. Money is just the tool you use to trade for that ability. It records wealth, but it is not wealth itself.

Assumption 2: debt is just owing money

A modern bank loan mostly creates money from nothing: lending you 100 does not move it out of someone else's deposit—it is written into existence the moment the loan is made, and destroyed when you repay (the Bank of England says exactly this). Which raises the classic suspicion: the bank created 100 but wants 110 back, so where does the extra 10 come from? Is this a Ponzi scheme waiting to blow up?

Half right, half wrong. The bank never has to print that 10: the same money changes hands many times a year, and when you borrow 100 to buy a machine that makes something worth 150, the extra value comes from labor, skill, energy, and demand—money just puts a price on it. So debt never borrows money; it borrows the future. Borrowing 100 today is a bet that you can produce more than 100 of real things later.

That is why the real danger is not "not enough cash to repay." Global debt reached roughly $348tn by the end of 2025, but the other end of every debt is someone's asset—repay it all at once and those assets vanish too. The system breaks when cash flow and confidence snap together, which is close to what happened in 2008.

Assumption 3: growth is forever, and it runs on people

For thousands of years growth almost never lost, because it kept running the same chain: more people → more production → more consumption → more GDP. The whole credit system ultimately rests on there always being people to work, consume, pay taxes, and borrow. But if the birth rate stays below replacement (about 2.1 children per woman) for a long time—fewer workers and borrowers, but more retirees and welfare spending—does that chain start to loosen?

Don't rush the conclusion, because one variable can rewrite the whole chain—AI. We used to assume GDP ≈ people × productivity per person; if AI multiplies one person's output many times over, the formula becomes GDP ≈ people × (AI-amplified productivity). As long as the multiplier climbs faster than population falls, fewer people need not mean a weaker economy. So whether low birth rates break the system is an open question with no answer yet.

AI is two sides of one coin. Zoomed out, it may rescue growth; zoomed in, it tears open a gap: when one person can do the work of twenty, the other nineteen jobs disappear at the same moment. The pie has not shrunk, but the number of people who get a slice—who have income—has. The formula can save growth but not distribution, and it pushes an old question, buried under growth for centuries, back to center stage: how should wealth be divided?

What counts as a real asset now

Across history, the "unit of wealth" keeps upgrading: grain → gold → currency → credit → data → and maybe AI next. Each upgrade pushes the carrier of wealth one step closer to productive capacity itself. Follow that direction and future wealth may increasingly equal one thing: how much future productive capacity you can control.

Feel it another way: $10bn in cash versus a 100GW power plant, a million GPUs, or a million AI agents—which is worth more? Many AI companies have already answered with their actions; they are fighting over compute, not gold. Energy, compute, robots, AI, knowledge, organizational ability, trust, attention—these are the underlying assets. Money is just the tool that keeps their accounts afterward.

But there is a twist: if AI, robots, and energy all become nearly unlimited, the last truly scarce thing may circle back to people—not people as labor, but human experience, creation, trust, attention, and feeling. AI can copy almost everything, except actually living a life in your place.

A posture, not a prediction

I am not going to hand over a clean verdict that future wealth will definitely be energy or compute. Two kinds of things live in this essay and should be kept apart: some are facts (how money is created, how debt circulates, population trends), and some are my extrapolation (that future wealth may lean toward energy, compute, or AI). Blending the two is exactly why so much "future prediction" fails to survive time.

And what I care about was never "will it blow up," but "what new system will it become." Institutions rarely wait for a full collapse before changing; more often they adjust the rules bit by bit—through inflation, debt restructuring, taxes, welfare, and monetary policy—slowly redistributing the pressure. First principles is not a prediction machine. It is a posture: take something you thought you understood, break it down until it cannot be broken further, and ask what is left.