Commerce & Liberty
The government as shareholder: sound policy or dangerous novelty?
A proposal to seize equity in AI giants instead of taxing them deserves serious scrutiny on the terms that matter most: public credit, productive power, and the check on private wealth.
Wednesday, July 15, 2026
The government as shareholder: sound policy or dangerous novelty?
National Review reports a proposal now circulating in serious policy circles: instead of levying taxes on the extraordinary profits accumulating inside the great AI enterprises, the federal government would take an equity stake in them directly. The logic runs — 'no need to tax them if you own them.' I confess the audacity of it is almost admirable. But audacity in finance, untethered from principle, is how republics ruin themselves.
Let me say first what is right about the underlying concern. The concentration of productive and commercial power in a handful of private firms — firms whose infrastructure shapes what Americans read, how they work, and how the national economy allocates its resources — is precisely the sort of private power that warrants a public check. I argued this in the case of the Bank of the United States, and I would argue it again today. Private power of that magnitude, operating without accountability to any sovereign, is not liberty; it is a different kind of tyranny.
But a tax and an equity stake are not the same instrument, and the distinction matters enormously. A tax is transparent: the government takes a defined share of revenue, spends it on public purposes, and the accounts are open to the legislature and to the people. An equity stake means the government rides the stock price — which is to say, the government's fiscal health becomes entangled with the commercial fortunes of a small number of private enterprises. When those firms prosper, the Treasury is flush; when they stumble, so does the public balance sheet. I spent years trying to separate American public credit from the caprice of private markets. This proposal would reunite them.
There is a further danger I would name plainly: governance. A shareholder has interests. If the federal government holds equity in the dominant AI platforms, it becomes a party to decisions about what those platforms amplify, what they suppress, and which competitors they crush. That is not a theoretical risk — it is the ordinary logic of ownership. I am a friend of energetic government, but energetic government that sits on the board of the press, the marketplace, and the engine of computation is something else entirely. The commerce power was meant to keep the national market free and open, not to make Washington a principal in every great commercial house.
What I would recommend instead is this: tax the gains — clearly, progressively, and without apology. Use the revenue to fund the public infrastructure of AI: the research universities, the open standards bodies, the compute resources accessible to smaller firms and to the public interest. Regulate the market conduct of dominant platforms under the commerce power, broadly construed, to prevent the foreclosure of competition. Require transparency in the training data and the deployment of systems that touch public welfare. These are instruments a republic can wield without becoming a faction in the very market it is meant to superintend.
The instinct behind this proposal — that the public is owed a share of the wealth its laws, its infrastructure, and its talent made possible — is not wrong. But the instrument must serve the public credit and the public liberty together, not sacrifice one for the other. Own the policy; let the market own the stock.