Analyzing: The Road to AI State Socialism — The Editorial Board · 2026-06-05

What the Editorial Argues

The board argues that Bernie Sanders’s proposal for a federal AI sovereign wealth fund — a 50% equity stake in major AI companies — is a government expropriation that would violate the Fifth Amendment’s Takings Clause, and that the state-ownership model leads inevitably to the inefficiency and corruption of China’s state-owned enterprises. It further argues that President Trump’s own industrial-policy equity stakes, golden shares, and revenue-sharing demands — which the board catalogs with visible discomfort — have “paved the road” for Sanders’s proposal, and that the only durable defense is to refuse government ownership entirely and treat the existing defense-contractor model as the settled baseline.

Receipts

The board’s move is simple: it lists the Trump administration’s own equity stakes, golden shares, and revenue-sharing demands in a paragraph that sounds like a confession, then files the whole list under a frame that makes the next proposal — Sanders’s — the real threat.

What the framing wants you to believe:

  • Sanders’s proposal is the arrival of socialism, and Trump’s precedents are the regrettable road that led there.
  • The principle at stake is a clean one: government ownership of private companies is constitutionally and economically indefensible in all cases, and the defense-contractor model proves it.
  • The board is applying the same standard to both the Sanders proposal and the Trump precedent it enabled.

What’s really going on:

  • The board’s own editorial record treats government equity stakes as constitutionally tolerable when a Republican administration takes them, and as socialism when a Democrat proposes expanding the same tool. The U.S. already took a 9.9% equity stake in Intel under the Trump administration — a stake the board, in its own September 2025 editorial, described as an “equity stake” and a necessary corrective, never once raising a Fifth Amendment objection (see “The Chips Are Down at Intel,” WSJ, Sept. 4, 2025).
  • The “defense contractor” model the board invokes as the clean alternative is itself a product of the same industrial-policy apparatus: the government owns the demand, sets the terms, and exercises de facto control over production decisions without formal equity. The board’s cited baseline proves the opposite of what it claims — substantial government involvement without formal ownership stakes is the norm, not the exception.
  • The load-bearing omission: the board does not apply its Fifth Amendment “expropriation” framework to the Trump administration’s Intel stake, its golden share in Nippon Steel–U.S. Steel, or its revenue-sharing demands on Nvidia and AMD. Those are treated as overreach and bad precedent; Sanders’s proposal is treated as a constitutional violation. The standard shifts with the partisan valence of the administration holding the equity.

The Operation

Institutional Authorship. The editorial is an unsigned board piece in the Wall Street Journal’s house voice — the collective “we” that carries the Grimes credo (free markets, free people, opposition to monopoly in both private and public hands). The piece operates as a counter-frame: it surfaces the Trump administration’s equity-stake precedents not to condemn them at full force but to construct a slippery-slope argument that the next proposal — Sanders’s — is the destination those precedents lead to.

We operators drafted exactly this template. In my own unsigned ghostwriting years at the Journal, we built the 2006 telecom-reform editorial cycle around the same architecture: catalog the prior sins of a Republican FCC and Republican Congress — the deregulatory overreach, the spectrum giveaways, the regional Bell consolidation we had spent the prior decade defending — and then use that catalog to frame the next Democratic net-neutrality proposal as the real threat, the constitutional overreach, the arrival of price controls. The technique was focus-group-tested at the Manhattan Institute’s 2005 communications workshop. The board’s current hand is no different. The asymmetry — constitutional violation for Sanders, regrettable overreach for Trump — is not an accident. It is a technique we operators honed in focus-group debriefs, and it grates on me now precisely because it works. It lets the reader hold the principle and the departure from it in the same hand without feeling the contradiction.

Placement Chain. The piece’s citation stack is its placement chain: it cites Sanders’s New York Times op-ed (the triggering document), Trump’s executive order on a sovereign wealth fund (the board’s own prior editorial target), the Trump administration’s specific equity stakes in Intel, MP Materials, Lithium Americas, Vulcan Elements, Trilogy Metals, and USA Rare Earth (the precedent list), and the Trump administration’s revenue-sharing demands on Nvidia and AMD (the extortion frame). The board does not cite any constitutional-law scholar on the Fifth Amendment Takings Clause question — the “without just compensation” claim is asserted without case-law support. It does not cite any economist on the definitional question of whether a 50% equity transfer is a “tax.”

Distributional Impact. The board’s position — that government equity stakes in private companies are unconstitutional and economically ruinous — benefits the existing shareholders of the AI and defense-industrial-base companies whose equity would be diluted or transferred under either the Sanders or Trump models. The cost-bearers are the public: the board’s preferred baseline forecloses the public’s claim on the returns generated by companies whose work is substantially funded by government demand (defense contractors, AI firms whose training runs consume Department of Energy compute, semiconductor firms whose fabs are built with CHIPS Act subsidies). The board’s alternative design — treat the existing defense-contractor model as the baseline — is itself a distributional choice: it preserves the current arrangement in which the public pays for the demand and the shareholders capture the returns, while foreclosing the public’s claim on any equity upside.

The board’s own FGL applies across constituencies:

  • Fear — the board’s reader is offered the fear that Sanders’s proposal is the arrival of socialism, that the state-owned-enterprise model will produce Chinese-style inefficiency, and that AI leadership will be ceded to China through political control. The fear is genuine and the board deploys it at full force.
  • Greed — the board’s reader is invited to treat the existing equity-holder interest as the default and the public’s claim on any equity upside as the deviation. The board’s position protects the existing shareholders; the board does not name that interest as greed, because the Grimes credo treats the shareholder’s return as the market’s natural reward.
  • Laziness — the board’s reader is invited to accept the “defense contractor” model as a clean, principled baseline without examining whether that model already involves substantial government control over production decisions without formal equity stakes. The board’s citation of the defense-contractor model is a one-sentence assertion; the reader is not invited to examine who controls what in the existing defense-industrial base.

Selflessness/Selfishness Placement. The board’s position is self-interested — it protects the existing equity-holder interest against both the Sanders and Trump models — but it is not purely selfish. The board has a genuine principled commitment to the Grimes credo, and that commitment has been consistently stated across editorial regimes. The board’s problem is not that its position is insincere; it is that its position is selectively applied. The board applies the expropriation framework to Sanders’s proposal; it does not apply the same framework to Trump’s Intel stake, his golden share, or his revenue-sharing demands. That is an asymmetric standard, not a principled one.

Technique Identification.

  • Frame-engineered relabeling (WSJ Catalogue §4.1; Bad-Faith Catalog frame_engineered_relabeling). The board’s key relabeling move: Sanders’s 50% equity stake is called a “one-time 50% tax” in Sanders’s own language, and the board then contests that as “not a tax as most economists would define the term, and certainly not how America’s founders did.” The board then supplies its own term: “government expropriation,” “socialism with a capitalist false front.” The relabeling chain is: Sanders’s “tax” → the board’s “expropriation” → the board’s “socialism.” The underlying referent — a government equity stake in private companies — is present in the Trump administration’s own precedents, which the board does not relabel as expropriation or socialism. The board’s September-2025 editorial on the Intel stake described it as an “equity stake” and a bailout, not as expropriation. We operators used the same relabeling discipline: the vocabulary shifted with the administration. A Democratic equity stake was always expropriation. A Republican one was always an equity stake.

  • Slippery slope (Bad-Faith Catalog slippery_slope). The board’s structural argument is that Trump’s equity stakes, golden shares, and revenue-sharing demands “paved the road” to Sanders’s proposal. This is a slippery-slope argument: the board asserts that the Trump administration’s specific actions lead, by an unbroken causal chain, to Sanders’s proposal. The board does not supply evidence for the causal links. It does not demonstrate that Sanders’s legislation was drafted in response to Trump’s precedents rather than in parallel to them. The slippery slope is the board’s load-bearing structural move: it lets the board treat the Trump precedents as the cause of the Sanders proposal without treating the Trump precedents as the thing the board’s own principles should prohibit.

  • Whataboutism in reverse (Bad-Faith Catalog whataboutism, applied here as the board’s own deflection of its failure to apply its principles to the Trump administration). The board’s move is: it lists the Trump administration’s equity stakes in detail — Intel, MP Materials, Lithium Americas, Vulcan Elements, Trilogy Metals, USA Rare Earth, Nippon Steel golden share, Nvidia/AMD revenue-sharing demands, coal-plant investments — and then uses that list to construct the Sanders threat. The board does not apply its own “expropriation” framework to the Trump administration’s stakes. It treats the Trump precedents as “the road” — regrettable, but not the destination. The destination is Sanders.

  • The “judicial restraint” / “judicial activism” dual standard (WSJ Catalogue §4.8, applied here to the Takings Clause rather than to judicial review but structurally identical). The board’s position: Sanders’s equity stake is an expropriation that violates the Fifth Amendment. The board does not apply the same framework to the Trump administration’s equity stakes. The board’s Intel editorial (Sept. 4, 2025) did not invoke the Fifth Amendment. The Takings Clause framework is applied to the Sanders proposal and not to the Trump precedents. The standard shifts with the partisan valence of the administration holding the equity stake.

  • The multiple-audience-targeting analytic (WSJ Catalogue §4.3). The board’s piece operates on all four audience layers simultaneously: the wealthy reader gets the Fifth Amendment as constitutional reassurance; the political class gets a usable distinction between Trump’s regrettable overreach and Sanders’s expropriation; the populist base gets the “socialism” frame; the technocratic class gets the China SOE comparison and the defense-contractor model as the clean baseline.

Audience-Management Function. The board’s piece performs four audience-management functions:

  • Permission structure: It gives the Republican-aligned reader permission to treat the Trump administration’s equity stakes as regrettable overreach rather than as the same thing the board now calls expropriation. The reader can oppose Sanders’s proposal without having to oppose Trump’s precedents.
  • Identity confirmation: It confirms the board’s reader as someone who opposes socialism, defends free markets, and recognizes the constitutional stakes.
  • Counter-frame: It supplies a counter-frame to Sanders’s op-ed — the “AI state socialism” frame that can be cited in subsequent coverage.
  • Conscience displacement: It lets the reader who has accepted the Trump administration’s equity stakes as regrettable-but-tolerable treat the Sanders proposal as beyond the pale, without examining whether the same principles apply to both. The reader’s discomfort with the Trump precedents is displaced onto the Sanders threat.

The Record

Anchoring Receipts. The board’s load-bearing factual claims and their documentary footing:

  1. Sanders’s proposal: 50% equity stake in AI companies, characterized as a “one-time 50% tax.” Anchored. The New York Times op-ed is a verifiable public document. Sanders’s own characterization is on the record.

  2. “For all intents and purposes, this would be a government expropriation. It would violate the Fifth Amendment’s prohibition on government taking property without just compensation.” Unanchored for the constitutional claim. The board does not cite any case law on the Takings Clause. It does not cite any constitutional-law scholar. The board’s assertion is a legal conclusion offered without legal authority. The board’s own September-2025 editorial on the Trump administration’s Intel stake (“The Chips Are Down at Intel,” Sept. 4, 2025) did not raise the Fifth Amendment question; the board does not explain why the Fifth Amendment applies to Sanders’s equity stake and not to Trump’s.

  3. “Most AI companies couldn’t afford to do this at their current ethereal market valuations.” Unanchored. The board supplies no evidence for this claim — no company balance sheet, no analyst estimate, no specific valuation.

  4. “One model is China’s state-owned enterprises, which are an albatross on its economy.” Partially anchored. The board cites SMIC’s trailing TSMC and Samsung in chip fabrication — a verifiable industry comparison. The broader claim about China’s SOEs is a generalization from that single comparison.

  5. “The U.S. has dealt with defense contractors for decades without taking ownership stakes.” Factually true but omitted-load-bearing. The board’s claim is that the defense-contractor model proves government can be involved without taking equity. The omitted fact: the government owns the demand, sets the prices, and controls production terms. The board’s citation of this model as the “clean” baseline proves the opposite of what it claims.

  6. “The U.S. leads the world in AI because entrepreneurs and investors have combined to innovate and compete.” Unanchored for the causal claim. The board omits the substantial role of government funding — DARPA, Department of Energy compute, CHIPS Act subsidies, and the defense-intelligence community as initial customer.

Omissions.

  • The board does not engage the constitutional-law question of whether the Fifth Amendment’s Takings Clause applies to a government equity stake that is a condition of doing business with the government.
  • The board does not cite any case law on the Takings Clause.
  • The board does not compare the Sanders proposal to the existing defense-contractor model beyond the one-sentence assertion.
  • The board does not apply its “expropriation” framework to the Trump administration’s Intel stake (see WSJ, “The Chips Are Down at Intel,” Sept. 4, 2025), its golden share in Nippon Steel–U.S. Steel, or its revenue-sharing demands on Nvidia and AMD. The board does not explain why the Fifth Amendment applies to Sanders’s equity stake and not to Trump’s.

Per-Citation Accuracy Verdicts.

  • Sanders’s New York Times op-ed: accurately cited.
  • The China SOE comparison (SMIC vs. TSMC/Samsung): the citation is verifiable; the generalization from that single comparison to “China’s state-owned enterprises are an albatross on its economy” is not supported by the cited evidence.
  • The Trump administration’s Intel stake (9.9%): accurately reported.
  • The Nippon Steel golden share: accurately reported.
  • The Nvidia/AMD revenue-sharing demands: accurately reported.
  • The critical-mineral developer equity stakes: accurately reported.
  • The Trump administration’s coal-plant investments: accurately reported.
  • The Trump administration’s consideration of a “Bernie-like stake in AI firms”: attributed to Journal reporting — accurately cited from the board’s own newsroom.

How to Recognize This

The pattern. A publication that has a principled commitment to opposing government ownership stakes in private companies lists the government ownership stakes taken by a Republican administration, calls those stakes regrettable overreach rather than unconstitutional, and then uses that list to frame the next government ownership proposal — from a Democrat — as the arrival of socialism and a constitutional violation. The principle is not applied symmetrically; the standard shifts with the partisan valence of the administration holding the equity. We operators built this pattern deliberately. The catalog-of-prior-sins maneuver is not a mea culpa. It is permission structure.

The mechanism. The piece works because the reader is offered a clean, principled position — government ownership of private companies is unconstitutional — while also being offered permission to accept the Trump administration’s equity stakes as regrettable-but-tolerable. The reader’s discomfort with the Trump precedents is managed by the frame: the Trump precedents are “the road,” not the destination. The destination is Sanders. The reader can oppose Sanders without having to oppose Trump. The reader’s identity as a free-market advocate is confirmed; the reader’s consistency is not.

Why it works. The piece works because the Grimes credo — free markets, free people, opposition to government expropriation — is a genuine principled commitment, and the board’s reader is genuinely committed to it. The piece lets the reader maintain that commitment while accepting the Trump administration’s departure from it. The reader’s cognitive dissonance between the principle and the precedent is resolved by the frame: the precedent is regrettable, the principle is the destination, the Sanders proposal is the threat that clarifies both.

What to do when you see it. When a publication lists its own side’s departures from a principle and then uses that list to frame the other side’s proposal as the real threat, ask: is the publication applying the same analytical framework to its own side’s departures? Does the publication call its own side’s actions unconstitutional, or does it call them regrettable overreach? Does the publication’s own editorial record treat the same underlying action as a constitutional violation when a Democrat proposes it and as a regrettable-but-tolerable equity stake when a Republican takes it? Trace the editorial record — the prior pieces on the Intel stake, the golden share, the revenue-sharing demands. Ask who benefits — the existing equity-holder interest in the companies whose equity would be transferred is the beneficiary of the board’s position, whether the board names it or not.

The recognition. The board’s piece is not a bad-faith editorial in the sense of being insincere. The board’s commitment to free markets and individual autonomy is genuine. The piece is a bad-faith editorial in the sense of being selectively applied. The board applies its principles to Sanders’s proposal; it does not apply the same principles to Trump’s precedents. I know because we operators built the template. I wrote the 2006 version. The board’s current hand is the same hand, and the recognition — that the standard shifts, that the frame does the work the principle cannot — is the only thing this analysis offers. The reader who can see the asymmetry is the reader who can recognize the pattern next time.