The oligarchy is fine, your kids aren't: three July data points and the shape of the next decade

On 20 July 2026, the Assemblée Nationale voted to ban children under fifteen from opening social media accounts. The same weekend, the official entrusted with U.S. AI safety resigned three months into the job. Two days earlier, the Wall Street Journal reported that 556,850 people worldwide now sit above the $30 million mark, a 14.4% jump in a single year, the fastest expansion since 2017. These are not separate stories. They are three readings from the same instrument.

The instrument is the gap between the speed at which a tiny elite compounds its safety nets and the speed at which democratic societies can legislate guardrails for the rest of the population. One number moves with the patience of capital. The other moves with the tempo of a parliamentary term. Watching them drift apart is the story of the next decade.

The money already moved

A record 556,850 people globally now hold more than $30 million in net worth. That figure, reported on 19 July 2026 by the Wall Street Journal, is up 14.4% in 2025 alone. The pace matters more than the headline: this is the sharpest annual expansion of the millionaire cohort since 2017. The growth is concentrated in the United States, where roughly 40% of the world's ultra-high-net-worth individuals reside, but the geographic spread is widening, with notable gains in Asia and the Gulf.

A counter-narrative is easy to construct. Stock markets have recovered; housing in major centres has stabilised at high levels; currency effects inflate dollar-denominated counts. None of that is wrong, and all of it is true. But the dollar argument explains growth, not pace. Fourteen-point-four percent in a single year, after a decade of central-bank balance-sheet expansion, points to a structural shift in how wealth is being allocated, not merely priced.

The speed of the state

Consider the contrast. France's under-15 social-media ban took roughly two years of public agitation, a Senate commission report, and a Macron-era political gamble to clear the Assemblée Nationale. The bill still faces constitutional review and possible referral to the Conseil constitutionnel. Implementation will likely require age-verification infrastructure that does not yet exist at national scale.

The resignation of the U.S. head of AI safety on 20 July 2026, after only three months in the role, is a different kind of lag. There the delay is not legislative but administrative: a position is created, a person is appointed, and within a quarter they walk out the door. The policy machinery either loses its operator or, more pointedly, loses the public signal that the position matters. Either reading is bad.

Both moves share a feature: each was supposed to be the answer to a problem the market had already monetised. France is legislating for a generation that has already grown up inside algorithmic feeds. America had to staff a regulator for systems already embedded in credit scoring, hiring, and policing. The state arrives after the bill has come due.

What a guardrail is worth

There is a defensible case that the French bill is the right instrument at the right moment. Cognitive development research on adolescent attention and reward systems is robust enough to survive constitutional scrutiny in most European systems. The bill also performs a political function: it gives parents and educators a public marker, even if enforcement proves uneven. Critics, including major platforms and digital-rights groups, counter that age-verification will inevitably push minors into less-moderated spaces, that it will entrench the identity-document industry, and that it will do little about the algorithm itself, which remains the proximate cause of compulsive use. Both points are real. The strongest version of the case for the ban does not deny them; it argues that imperfect legislation now beats perfect legislation in five years, by which point another cohort has graduated through the system.

On AI safety, the harder counter-narrative is that the resignation is not a setback at all but a correction: the role may have been designed as a fig leaf, and a competent official is right to leave it. A more cynical reading is that the position is being deliberately starved, and the resignation is the predictable outcome of a brief in which no one expects the appointee to succeed.

The pattern underneath

These threads converge on a structural pattern that does not need a label from an academic theorist to be legible. The platform economy monetises attention. Attention, in minors, is a future-revenue stream; in adults, it is the present-margin stream. The state, when it acts, regulates the surface, age, screen time, content categories, while leaving the underlying business model intact. Capital, by contrast, adjusts without friction: it reallocates to the jurisdictions with the lightest guardrails, prices the regulatory risk into valuations, and continues to compound.

The asymmetry is not new. What is new is its acceleration. A 14.4% annual expansion in the ultra-wealthy cohort is not, on its own, a scandal. It becomes one when set against the political time required to draft, pass, and implement a child-safety bill, or to keep a single AI-safety official in office for longer than a quarter. The wealthy can wait. The minors being legislated for cannot.

What to watch

Two dates will be informative. The first is the Conseil constitutionnel's ruling on the French bill, expected in the autumn legislative window. The second is whether the U.S. AI safety role is filled, and on what timeline, after the 20 July resignation. A vacancy that persists past the autumn would tell its own story about the durability of the post.

What the available reporting does not yet settle is whether the under-15 ban survives contact with implementation. France has a strong record of passing symbolic tech legislation and a weaker record of enforcing it at scale, a pattern observers of the CNIL and Hadopi will recognise. The millionaire count, by contrast, is not contested; it is the baseline against which everything else will be measured.

Desk note: Monexus ran these three wires together because each one, taken alone, reads as a discrete policy event. Together they describe a single political-economic condition: the velocity of capital versus the velocity of the state, measured in a single news cycle.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://t.me/polymarket/21209
• https://t.me/polymarket/21203
• https://t.me/unusual_whales/17882
• https://t.me/unusual_whales/17879
• https://t.me/polymarket/21188