The standard explanations for gold's 55% surge in 2025 are well-rehearsed: central bank de-dollarization, geopolitical hedging, AI bubble insurance, inflation expectations. All valid. None sufficient. There is a deeper, structural force at work that the market has not yet named.

The world's largest asset class is being destroyed.

Not equities. Not bonds. Not real estate. Human capital — the aggregate economic value of skills, knowledge, training, and experience embedded in the global workforce. The World Bank estimated it at $1,143 trillion in 2018, roughly four times all produced capital and fifteen times all natural capital. The entire architecture of modern economic life — mortgages, pensions, education debt, consumer spending — is predicated on the assumption that this asset retains its value over a working lifetime.

Artificial intelligence is breaking that assumption. And the money is looking for somewhere to go.

Devaluation, Not Disruption

Every prior technological revolution left an upper frontier of human capability intact. The steam engine replaced muscle but not management. The computer automated arithmetic but enhanced analysis, design, and strategy. The disruption was horizontal — it moved across industries — but never vertical. It never reached into reasoning, judgment, and expertise.

AI is vertical. A $20/month subscription now produces competent legal analysis, working software, financial models, architectural drafts, and research synthesis — in minutes. These are not factory-floor tasks. These are the cognitive skills that decades of education and professional experience were supposed to make uniquely valuable. The quality is imperfect but improving faster than any human learning curve.

What this does to human capital is not what disruption does to an industry. It is what inflation does to a currency. It does not move value. It destroys it. The skills still exist, but what they command in the market is collapsing. Goldman Sachs strategists recently called it the "commoditization of expertise." The term deserves broader application: when a generalist AI can replicate the output of a specialist human, the scarcity premium on expertise evaporates. The people who spent decades building their human capital are watching its market value decline in real time, like homeowners in 2008 watching their equity vanish — except there is no Fed put for skills.

The Missing Thesis

Here is the argument nobody is making: when the primary asset most humans hold — their earning capacity — is being devalued, the rational response is to convert current income into assets that cannot be commoditized. Gold cannot be replicated by an algorithm. Its supply is constrained by geology, not innovation curves. It carries no counterparty risk and requires no productive economy to retain value.

This thesis does not require investors to consciously articulate it. Markets aggregate unspoken anxiety efficiently. When a 45-year-old financial analyst watches AI produce in seconds what took her team a week, she does not think "I should buy gold." She thinks: "What is my future income worth?" That question, multiplied across millions of knowledge workers worldwide, manifests as a structural rotation out of faith in human-capital-derived income and into hard stores of value.

How the Value Transfers

Human capital is not liquid. You cannot sell your skills on an exchange. The transfer to gold operates through four reinforcing channels:

The Individual Hedge. A professional who perceives her earning power as a depreciating asset cannot sell that asset — human capital is not traded on any exchange. But she can accelerate the extraction of value from it while the market still prices it highly. The most direct route is redirecting current income — salary, fees, billings — toward hard assets at peak earning levels. But there are less obvious routes: borrowing against future earnings while lenders still underwrite them at today's rates, effectively shorting your own career trajectory; selling a practice or consultancy while clients still pay a premium for human-delivered expertise; converting equity compensation into hard assets before AI-driven margin compression hits the employer's valuation. Each of these is a window that closes as AI advances. The professional who acts early is not speculating on gold. She is extracting residual value from a depreciating asset before the depreciation becomes visible to the market.

The Corporate-to-Capital Channel. When a company replaces analysts with AI, labor costs fall but output does not. The surplus — previously captured by workers as wages — flows to shareholders as profit. Those shareholders, newly enriched, understand that AI-driven margin expansion is simultaneously destabilizing consumer spending. The rational portfolio response: diversify away from assets that depend on mass wage income and toward assets that do not. Gold and land qualify. Consumer equities do not.

The Institutional Channel. Pension funds calculate liabilities based on projections of future wages. If AI compresses wages across professional classes, contributions decline while liabilities remain. The funding gap widens. Institutional investors do not need to articulate an "AI devalues human capital" thesis. They simply observe that their wage-growth assumptions are failing and adjust allocations. Central banks are further ahead: their record gold accumulation — over 1,000 tonnes annually for three consecutive years — may already reflect an unstated awareness that the tax-and-spend model depends on a wage base that is under structural threat. China, the most aggressive buyer, is also the sovereign most exposed to AI-driven automation. This may not be a coincidence.

The Fiscal Channel. When AI compresses wages across professional classes, tax revenue declines and social expenditure rises simultaneously. The arithmetic is merciless: whether the policy response is called UBI, retraining subsidies, or stimulus, it must be funded. In a world where productivity gains accrue to capital owners, the tax base narrows as spending demands widen. The escape valve is monetary expansion. And monetary expansion is the most reliable long-term driver of gold. Each link in this chain — AI replaces cognitive labor, governments print to support displaced workers, currency debasement drives gold — has extensive historical precedent.

These channels are not independent. They form a reinforcing loop: individual hedging reduces consumer spending, which weakens revenues, which accelerates AI adoption to cut costs, which displaces more workers, which increases fiscal pressure, which drives monetary expansion, which pushes more capital toward gold. The loop has no natural circuit-breaker as long as AI capabilities continue to improve.

The standard rebuttal is historical: every wave of automation created more jobs than it destroyed. But that argument assumes a frontier of human capability that technology cannot reach. The question AI poses — for the first time in economic history — is: what if there is no such frontier? What if the technology follows humans to every frontier they retreat to? If so, the implicit guarantee underwriting $1,100 trillion in human capital — that investing in education and experience would yield returns over a working lifetime — is void. The money has to go somewhere.

Gold is the most visible beneficiary, but not the only one. Any asset whose value derives from physical scarcity rather than human productivity stands to gain. Land cannot be generated by a neural network. Infrastructure requires physical construction regardless of how intelligent the software becomes. Energy assets retain value because AI runs on electricity, not on good intentions.

The Arithmetic

Global human capital: $1,100+ trillion. Total above-ground gold at current prices: roughly $35 trillion. A one percent reallocation from one to the other — a rounding error in a $1,100 trillion asset class — represents $11 trillion of demand against a $35 trillion market. This is not a prediction. It is arithmetic. The point is not that such a reallocation happens overnight, but that the structural incentive to move wealth from depreciating assets to non-depreciating assets is now permanently in place, and intensifies with every improvement in AI capability.

The conventional gold thesis is cyclical: inflation, war, currency collapse. These come and go. The thesis outlined here is structural. It does not depend on any particular geopolitical scenario, interest rate path, or market correction. It depends on a single proposition: that AI is permanently reducing the economic value of human cognitive skills, and that the $1,100 trillion of human capital built on the assumption of durable skill value is being repriced.

If this proposition is even partially correct, then gold's current price reflects only the early stages of a repricing that could last decades. Not because central banks are buying, and not because the dollar is weakening, but because the most fundamental asset in the global economy is losing its value, and the wealth it once represented is looking for a new home.

Gold cannot think. That is now its greatest advantage.