A share levy into a stock-owning Trust Fund, capped at a third

The tax pieces of this system are claims on income. This piece is a claim on capital, and it exists because of a simple observation: in a strong-AI economy the return to owning machines rises while the return to working falls. A universal income funded only by taxing the shrinking wage base will shrink with it. A fund that owns a slice of the machines grows with them.

How the levy works

Each year, every large public company issues new non-voting shares equal to 2.5 percent of its market value and delivers them to the Trust Fund, with banks as custodians (piece 2). No cash changes hands, so the levy never squeezes a company's operations; it dilutes existing owners by 2.5 percent a year, which is the whole point. The shares do not vote. The fund is not an activist and does not sit on boards. It collects dividends and holds. Private companies above a size threshold would owe an equivalent levy in cash or shares; the model treats the public market as the base and private companies as a reserve.

The levy stops when the fund owns 33 percent of the market. Anything the fund holds above 30 percent is sold into foreign assets, so the fund's U.S. share oscillates between the two lines and never climbs. The model has it reaching 29 percent in 2040 and holding there.

Why the cap

Round seven of the funding math removed the cap to see what happened. The fund owned 64 percent of the U.S. market by 2045 and 96 percent by 2060. A fund that owns almost everything is the state owning almost everything, whatever the shares are called, and every decision about which company lives or dies would run through it. The cap at a third, with the valve, keeps the public a large minority owner and no more. The foreign assets are the steam valve: they let the fund keep growing (half the fund is abroad by 2045) without the domestic share creeping up. They also diversify the public's claim away from the U.S. market alone.

The draw

The fund pays nothing into the pool until 2036. From 2036 the draw ramps from 1 percent of the fund to 3 percent by 2046 and stays there. The delay matters: in the early years every dollar stays invested, which is why the fund reaches $39 trillion by 2035 and $129 trillion by 2045 on 7 percent U.S. returns and 5.5 percent abroad. By 2045 the draw is $3.7 trillion a year, about a sixth of the pool. The honest finding from the master model is that the fund never covers the full cost alone; it is one leg of five.

Pooled, not individual

The fund has no individual accounts and no inheritance. Everyone alive draws from the same pool; nobody's share passes to their children. This is the opposite of the "baby bonds" and private-account designs, and it is deliberate. Individual accounts recreate the inequality they were meant to fix (returns compound for the lucky), they need 262 million statements and a customer-service operation, and they turn a public claim on capital into a private one that can be borrowed against and lost. A pooled fund is a right of citizenship, like a vote, not an asset.

Risks we take seriously

Announcing a permanent 2.5 percent annual dilution will reprice the U.S. market on the day it is announced; the model's market-value path does not include that drop, and the levy's yield in the early years would be lower than shown. Foreign governments may object to a U.S. public fund buying their companies; the valve assumes they do not. And the 7 percent U.S. return is the historical average, not a promise. Cut it to 5 percent in the model above and the fund in 2045 is a third smaller. The design still works; it just takes longer.