The U.S. Government Owns 9.9% of Intel and Some Thirty Companies. It's Not Keeping Score.

By the standard count, just over a year ago the American government held equity stakes in zero private companies. Today it owns roughly a tenth of Intel, a stake in the country's only integrated rare-earth producer, a veto over U.S. Steel, a thread of equity through nine quantum start-ups and a facilitation role in an $80 billion nuclear build-out — about thirty firms in all, purchased with perhaps $27 billion of public money. It is the quickest accumulation of state equity in the American private sector in decades. The remarkable thing is not that it happened, but that almost nobody, including the government itself, can account for it.

The emblem is Intel. In August 2025 the federal government bought 433.3 million Intel shares at $20.47 apiece, roughly $8.9 billion in all, in a purchase tied to the CHIPS chip programme that left it with a holding of some nine per cent. Intel's stock has since multiplied several times over, lifting the stake to, at its peak this year, roughly $56 billion. That is an extraordinary paper profit for Washington. The trouble is that the profit appears in no budget, no filing, no public ledger. The United States has made a very large trade and, by the rules of federal accounting, is not permitted to show its result.

How a government became a shareholder

The mechanism is less exotic than it looks. Much of the portfolio was built out of the CHIPS and Science Act, the 2022 law that subsidised domestic chip investment. The Biden administration spent its research-and-development money as plain grants. The present administration has re-read the same appropriation to mean something else: to receive the R&D cash, a company must now hand back equity, warrants, or a share of its revenue. Nineteen awards totalling up to $3.8 billion have been tied to equity, spread across eighteen companies. On top sit the bigger, one-off deals — the Intel purchase, a $400 million block of MP Materials that makes the Pentagon its largest shareholder, a "golden share" over U.S. Steel kept when it was sold to Japan's Nippon, and a scattering of quantum and minerals positions, some still closer to term sheets than to signed deals.

It is not, importantly, a single portfolio. The stakes are scattered across at least four agencies — Commerce, the Pentagon, the Development Finance Corporation and the Energy Department — and there is no consolidated ledger. The most complete public accounting is a tracker kept by the Council on Foreign Relations, which notes that announced deals are "the tip of the iceberg." The reason the holdings are so hard to pin down is also the reason they are so striking: only one of the agencies involved, the DFC, has a clear statutory authority to buy equity, and that was written in 2018 to finance ports in developing countries. The rest of the portfolio rests on an interpretation of laws that authorise "grants and other transactions" but never say, in so many words, that the government may own the stock.

The arithmetic, and why it cannot be booked

The Intel position makes the paradox concrete. Buy at $20.47, hold a volatile turnaround name now trading above $100, and the $8.9 billion is worth several multiples of its cost. Yet under federal budget rules an equity purchase is recorded as an outlay the day it is made, with little mechanism for recognising a return. The gains are real in the market and invisible to the Treasury. The market can mark the position; Washington cannot.

There is a further oddity in the structure of the deal that tells you who is really in charge. The government's Intel stake is passive — no board seat, no information rights, and an agreement to vote in line with Intel's board. Most of the shares were delivered at closing, with a portion held in escrow to be released only as Intel meets the milestones of a Pentagon chip programme. The claw-back and profit-sharing terms that had attached to the earlier subsidy were dropped. In other words, the state gave up any guaranteed return in exchange for a slice of whatever the stock becomes. It is less a secured claim than an outright bet on an American chipmaker's redemption, one that has, so far, paid.

The four hats

This is where the investor has to slow down, because the same institution is wearing four hats at once. The government is Intel's regulator, a customer through the Pentagon chip programme, a financier, and now a shareholder. Those roles pull in different directions, and nothing is set up to resolve the tension. Unlike the 2008 bank rescues, this portfolio has no special inspector-general, no dedicated congressional panel and no routine audit.

The case for doing it is not frivolous. Private capital underinvests where the strategic value — national security, supply-chain sovereignty — exceeds the commercial return. A government equity stake aligns Washington's interest with the survival of a firm the country cannot afford to lose, and it appears to do something money alone cannot: recruit private partners. SoftBank followed the government into Intel; Brookfield and Cameco into the nuclear build-out; and the government's presence, reporting suggests, helped pull Apple into a preliminary chip-making arrangement. A state backer on the cap table is, in a real if blunt sense, a de-risking. The same logic runs through rare earths: 72 per cent of America's supply still comes from China, and the Pentagon's bet on MP Materials is an attempt to move that number.

The case against is equally plain. A government that is simultaneously rule-maker, buyer, funder and owner distorts the very market it is meant to serve, and it invites the question of which firms are chosen and why. And because there is no exit strategy, the stakes do not unwind: they become permanent, and they pass to each incoming president as a fresh instrument of power. The administration calls it a "portfolio approach," a pseudo-sovereign wealth fund under executive control. The more accurate description is a shareholder state operating without a scoreboard.

A sovereign on the cap table

The practical consequence for an investor is twofold, and the two pull apart. First, a new category has appeared: the strategically essential company with a sovereign co-investor. MP Materials receives a price guarantee and an offtake agreement alongside its government stake; Intel receives a customer, a backer and a political narrative. If you own, or are watching, these names — chips, rare earths, nuclear, the quantum cluster — you are now investing beside a government that has decided where the industrial risk sits. That is, in itself, information.

Second, and this is the harder one, none of it is legible in the usual way. There is no consolidated disclosure, no TARP-style report card, and the government's "wins" are paper gains on volatile stocks it is not allowed to book. AInvest's aggregate signal, a composite of analyst ratings, labels Intel a Hold — a reminder that the state's spectacular mark sits on a contested turnaround, not a settled thing. The right posture, then, is not to read the government's entry as a seal of approval or a floor. It is to judge the business on its own merits, and to carry one new fact into every such position: a sovereign is on the cap table, it has no stated exit, and nobody in Washington is keeping score.