# American State Capitalism The name for what happened when the United States government stopped subsidizing strategic companies and started **owning** them. Since January 2025 the federal government has taken equity or quasi-equity positions in dozens of private firms — semiconductors, critical minerals, nuclear energy, quantum computing — in a break with roughly seventy years of postwar policy that used grants, loans, and tax incentives but avoided ownership. For the deal-by-deal ledger, see [[Government Equity Stakes Roster]]. ## The distinction that matters Washington has owned pieces of companies before. What is new is **why**, and **for how long**. Every significant precedent — the Reconstruction Finance Corporation, the Chrysler warrants, Continental Illinois, [[Troubled Asset Relief Program]], the CARES Act airline warrants — was a **rescue**: reactive, crisis-triggered, and built to exit. TARP had a statutory sunset. The Chrysler warrants were sold at a profit in 1983. The RFC filed a final report and closed. The current program is the inverse. It is **proactive** rather than reactive — the targets are not failing companies but companies deemed strategically necessary — and it carries **no sunset, no holding period, and no disposition rule**. That is the whole conceptual difference. A temporary stake is a bailout. An open-ended one is an ownership class. ## The stated rationale The administration's case, and the reason it draws support well outside MAGA: - **China dependency is real.** The U.S. imports roughly 72% of its rare earth elements from China, which controls over 90% of global refining. Over 90% of the most advanced semiconductors are made in Taiwan. These are genuine chokepoints, not invented ones. - **Some sectors will not be built by profit-seeking capital alone**, because the competitors are state-backed and loss-absorbing. Patient capital has to come from somewhere. - **Taxpayers who take the risk should share the upside.** This argument is why [[Bernie Sanders]] backed the Intel stake, and why Sanders and Elizabeth Warren tried to *require* equity in the 2022 [[CHIPS and Science Act]] — an amendment the Senate rejected. Taking the rationale seriously is what makes the critique sharp rather than reflexive. The problem is not that the state invested. It is *how*, and with what constraints. ## The four failure modes **1. Politicization of corporate decisionmaking.** An owner can steer. The Intel deal was structured as passive — no board seat, government votes with the board — but structure is a promise, not a lock, and promises are revised. **2. Investor versus regulator.** [[Department of Commerce]] holds Intel stock *and* writes the semiconductor export-control rules that determine which of Intel's competitors can sell what, where. There is no wall between those roles and none has been proposed. **3. Cronyism, structurally unfalsifiable.** Nothing published explains why these firms and not their competitors. Absent a disclosed selection process, "strategic merit" and "political connection" produce identical observable outcomes. The [[Vulcan Elements OSC $620M Pentagon Loan]] — where ProPublica documented White House initiation of a loan benefiting a firm partly held by a [[Donald Trump Jr.]]-linked fund — is the proof of concept that the thumb does reach the scale. **4. Approval as currency.** The [[NVIDIA]] and [[AMD]] China arrangement takes a reported 15% of revenue from chip sales to China in exchange for export licenses. The [[US Steel Nippon Golden Share]] took a perpetual governance veto as the price of [[CFIUS]] clearance. Neither required an appropriation. Both converted a **regulatory decision the government was going to make anyway** into a financial or governance claim on a private firm. This is the mode with the fewest checks on it, and it is the one most likely to generalize. ## The accountability hole The positions **appear in no budget document and are overseen by no dedicated watchdog.** There is no federal registry. Independent trackers at CFR, CSIS, and several newsrooms produce materially different totals — $10B to $27.6B, 30 to 37 deals — not because anyone is wrong but because the underlying disclosure does not exist and each has to draw the equity/quasi-equity line itself. A government that owns dozens of companies and cannot say so in a budget line is not running an industrial policy. It is running a portfolio off the books. ## Why the courts will not settle this Every stake so far was taken **with the company's formal consent**, which forecloses the obvious lawsuit — a firm that just handed over a tenth of itself will not sue. The plaintiffs with real standing (a rejected applicant, a diluted shareholder, an undercut competitor) have not appeared. As Peter Harrell put it in Lawfare, if Congress wants oversight of this "new age of American state capitalism, it will need to act to assert its authority — rather than waiting for the courts." Congress has not acted. ## Connections - The ledger: [[Government Equity Stakes Roster]] - The instrument with no economic stake: [[Golden Share]] - Where this ends up if it is institutionalized: [[Sovereign Wealth Fund]] - The sector driving most of it: [[Critical Minerals]] - The corruption case study: [[Vulcan Elements OSC $620M Pentagon Loan]] ## See Also - [[Government Equity Stakes Roster]] - [[Golden Share]] - [[Sovereign Wealth Fund]] - [[CHIPS and Science Act]] - [[Defense Production Act]] - [[Critical Minerals]]