The Scissors
Programmed inequality: How wealth distribution is completing a century-long rotation — and why artificial intelligence is accelerating its closure.
March 2026. The World Inequality Report captures the snapshot: the richest 10% of the planet owns three-quarters of all global wealth. The poorest 50% holds 2%. Two days ago, Larry Fink — CEO of BlackRock, fourteen trillion dollars under management — wrote in his annual letter that artificial intelligence risks replicating this pattern “on an even larger scale.”
The man who manages more capital than anyone else is warning his shareholders that the mechanism enriching them could break the system that makes it possible. The warning arrives as data documents that the global distributive structure is completing a 360-degree turn: a return to the configuration of the Roaring Twenties, when coal and steel magnates sat atop a pyramid nearly identical to that of today’s lords of silicon and compute.
The scissors are reopening. And this time they have an accelerator.
The U-Curve
In 2016, economists Emmanuel Saez and Gabriel Zucman reconstructed wealth distribution in the United States from 1913. The shape that emerges is a sharp U. Concentration was high in the 1920s: the share of the top 0.1% reached 25% in 1929. Then it collapsed. It continued falling through the 1970s, in the context of the Great Depression, the New Deal, the war economy, and postwar welfare. The historic low came between 1970 and 1980.
That decade now appears as an anomaly. The bottom 90% of the population — the middle class, workers, small savers — held 35% of total wealth. The top 1% hovered between 23% and 25%. It was the antithesis of the current model.
This equilibrium was not a market miracle. It was the result of a precise architecture.
Aggressive taxation on large fortunes, marginal rates above 70%, unions capable of negotiating the distribution of value. Labor managed to hold its own against financial returns. The World Inequality Report 2026 shows that this interlude was not only American. In Europe, the postwar period produced the same compression. But it also documents that it has ended everywhere. In India, the income share of the top 10% went from 40% in 2000 to 58% in 2023. In South Africa, the most unequal country on the planet, the bottom 50% has negative net worth: more debt than assets. Since the 1980s, the climb back has been constant. The U is closing.
The Crossover – The Inequality Scissors
The late 1980s triggered a mutation. Deregulation policies on both sides of the Atlantic reversed the elite’s trajectory, which began a vertical climb. That of the mass of workers started a decline that has not yet stopped.
The moment the blades crossed again was the year 2000. That year, the top 1% (at 33%) surpassed the bottom 90% (stuck at 29%). It was the definitive victory of financial assets over wages. Capital stopped functioning as a tool for creating jobs and became an end in itself.
Federal Reserve data capture the result: in 2018, the top 1% held over 38% of total wealth. According to the OECD, the United States is the industrialized country with the highest share in the hands of the 1%: 40.5%.
The Circuit
Behind the curve lies a mechanism. Thomas Piketty made it famous: when the return on capital — profits, dividends, capital gains — consistently exceeds economic growth, wealth accumulates faster than labor can produce it. Those who own capital see their assets grow; those who depend on wages lose ground with every cycle.
The point is not that this happens occasionally. It’s that it doesn’t self-correct. On the contrary: it amplifies. Returns generate additional capital, which generates higher returns. It’s a positive feedback loop. Output is reinvested as input, and each iteration widens the gap between those inside the loop and those outside it.
Projections indicate inequality levels closer to the 19th century than to the postwar era.
Before the 20th century, this gap was wide enough to sustain the extreme concentration levels of the pre-industrial era — the world of Balzac, where inherited wealth determined destiny. During the 20th century, in an era catastrophic for capital but favorable to growth, the gap narrowed. With economic growth around 1.5% and capital returns trending toward their historical value of 4-5%, projections indicate it is reopening.
The Switch
If the circuit tends toward concentration, what interrupted it in the 20th century? The answer is not consoling.
The compression of inequality between 1930 and 1975 was not the product of capitalism’s natural progress. It was the collateral effect of catastrophes: two world wars, the Great Depression, the physical destruction of accumulated capital. The New Deal, the welfare state, progressive taxation — the architectures that produced that interlude — were born from the rubble. From the need to rebuild order after the previous one had disintegrated.
The loop was interrupted because capital was destroyed, not because someone found a way to regulate it. The gains of the 1970s were children of a social pact born from ashes, not a farsighted choice. The exception required conditions no one would want to replicate.
The Technological Pivot – Programmed Inequality
If in 1920 inequality was fueled by coal and railroad monopolies, in 2026 the new engine has a technical name: compute. Technology has become the pivot of the scissors. While it creates immense value, its ultra-concentrated ownership — data, chips, infrastructure — pushes the blades in opposite directions.
The mechanism operates through two channels. First: infrastructural concentration. Control of compute, training data, and physical infrastructure is centralized in a handful of global companies. The cost of training frontier models — hundreds of millions of dollars per cycle — creates barriers to entry that favor structural oligopoly. Unlike Standard Oil’s product monopoly, here the concentration concerns the very capacity to produce artificial intelligence.
Since 1989, a dollar invested in the stock market has grown fifteen times more than a dollar tied to the median wage.
Second: transfer from labor to capital. A paradox complicates the most common narrative: AI might reduce wage inequality, because it automates high-income tasks too. But it simultaneously increases wealth inequality, because the cost savings companies achieve through automation produce returns that flow to those who own the companies, not to those who worked there. Fink himself acknowledges this in his letter. AI risks replicating that pattern on an even larger scale.
Projections for 2026 indicate that the top 1% will rise to 39%, while the bottom 90% will fall to 22%. In 1920, the ratio was 45% versus 20%. We have digitized inequality, replacing industrial barons with cloud barons. But the effect on the ordinary citizen’s wallet is a plunge a hundred years into the past.
The Anchor
While the 1% and the 90% experience dramatic swings, there is a group that seems immune: the intermediate 9% — the upper bourgeoisie and the managerial class. This bracket has maintained an extraordinarily stable share: 35% in 1920, 42% in 1980, a projected 39% for 2026.
In a system where everything oscillates, a stable element operates as a regulator. The intermediate 9% represents the buffer: the managers, the elite professionals who run the system’s architecture on behalf of the top. The only group that has crossed the century without losing share, insulated from the crises that strike the base.
But projections signal a possible breaking point. For the first time, the 1% and the 9% would reach near-perfect parity: 39% each. The narrow elite is becoming as powerful as the entire class that supports it. If this convergence holds, the buffer loses function: there is no longer enough differential to justify the managerial class’s loyalty to the top. The system loses its internal stabilizer.
Programmed Inequality: Where To?
The World Inequality Report 2026 records that 60% of contemporary billionaire wealth comes from inheritance, monopoly power, or privileged connections — not from innovation. In the United States, 90% of children born in 1940 earned more than their parents; for those born in the 1980s, the probability has dropped to 50%. In a growing number of countries, the bottom 50% of the population owns less than nothing.
The Roaring Twenties ended with the crash of 1929 and a global catastrophe. The gains of the 1970s were born from that crisis’s ashes. The U-curve documented by Saez and Zucman describes a historical pattern, not a law of nature. Nothing guarantees the cycle will repeat — and nothing guarantees it will self-correct.
Infinite concentration is a physical impossibility. The question is whether we can imagine a correction that doesn’t require a catastrophe first. The last time the scissors were open like this, it took thirty years, two wars, and a financial collapse to close them. And those who closed them weren’t trying to. They were trying to survive.
Postscript
Larry Fink’s annual letter to BlackRock shareholders was published on March 25, 2026. It is the first time the world’s largest asset manager has included an explicit warning about AI’s systemic risks to wealth distribution.
