On Feb. 27, U.S. Secretary of Defense Pete Hegseth announced—via tweet—that he had directed his Department to “designate Anthropic a Supply-Chain Risk to National Security.” Accordingly, Hegseth claimed, “no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.” As one former Trump advisor put it, the designation amounted to “attempted corporate murder.” Soon thereafter, U.S. President Donald Trump chimed in via Truth Social, directing “EVERY Federal Agency in the United States Government to IMMEDIATELY CEASE all use of Anthropic’s technology.” “The Leftwing nut jobs at Anthropic,” the President threatened, “better get their act together, and be helpful” in phasing out their technology from government use or he would “use the Full Power of the Presidency to make them comply, with major civil and criminal consequences to follow.” Trump invoked no authority for his actions, although the Treasury Department, for one, soon followed his lead. The Department of Defense officially informed Anthropic of the designation on Mar. 5, and even before that official communication some defense contractors were already reportedly removing Anthropic’s frontier AI model Claude from their systems in response to the Administration’s perceived “blacklisting” of Anthropic. While there are reports that talks may be ongoing between the Department of Defense and Anthropic to attempt to resolve this dispute, the underlying issues raised by the Administration’s attack on Anthropic will persist.
The use of rarely-invoked national security authorities to target Anthropic is not a one-off action by this Administration. It should instead be viewed as the latest chapter in a concerted campaign of pretextual retaliations—parading in the guise of emergency national security regulation—that have characterized the second Trump Administration. Anthropic is but the latest in a long and growing list of targets of Donald Trump’s punitive presidency. President Trump has previously attempted to wield the power of the presidency to punish, among others, his political opponents, law firms that hired or represented them, universities whose administrative decisions he disagreed with, journalistic outlets that refused to use his preferred terminology, and companies that refused to fire Trump’s political enemies.
The background of the Trump Administration-Anthropic dispute has been treated in greater depth elsewhere. But the underlying disagreement is straightforward. The Department of Defense demanded that Anthropic, as a Department contractor, not prevent its technology from being used both for domestic mass surveillance of Americans and for fully autonomous lethal weapons. After Anthropic refused to abandon those core principles against such uses, Hegseth not only announced that the Department’s contract with Anthropic would be cancelled, but labeled Anthropic a “supply chain risk”—exiled, by fiat, from doing business with the federal government as well as (according to Hegseth) any other government contractor. What Anthropic believed to be core principled limitations on how its technology should be used, Hegseth characterized as Anthropic’s “cowardly act of corporate virtue-signaling that places Silicon Valley ideology above American lives.”
No Statutory Basis for the Designation
For the reasons that others have documented, Secretary Hegseth’s designation of Anthropic as a supply chain risk clearly exceeded his authority under the supply chain risk statute (10 U.S.C. § 3252) and Federal Acquisition Supply Chain Security Act (FASCSA, codified at 41 U.S.C. §§ 1321-1328 and 4713). In the first place, Anthropic does not meet the statutory definition of a “supply chain risk,” which is typically a designation for foreign companies that the U.S. government assesses to pose a national security risk. Under § 3252(3)(4), a supply chain risk is defined as “the risk that an adversary may sabotage, maliciously introduce unwanted function, or otherwise subvert the design, integrity, manufacturing, production, distribution, installation, operation, or maintenance of a [system] so as to surveil, deny, disrupt, or otherwise degrade the function, use, or operation of such system.” It is very hard to see how Anthropic’s refusal to contract its technology for mass surveillance or lethal autonomous weapons meets this statutory criterion. Only one company has ever been publicly designated as a supply chain risk (a Swiss company with alleged Russian ties that was barred, specifically, from intelligence community contracts); no domestic company has ever been targeted.
Moreover, the scope of the Anthropic designation is similarly both unprecedented and unlawful. No previous supply chain risk designation has ever sought to prevent all federal contractors from dealing with a targeted company on all matters: this is a government-compelled secondary boycott designed to stigmatize Anthropic as a corporate pariah. But such a secondary level action is unlawful: on its face, § 3252 applies only to federal contracts and subcontracts, and only when an agency head has determined that “less intrusive measures are not reasonably available to reduce [] supply chain risk.” The designation here is intentionally overbroad.
Moreover, none of the procedural requirements of § 3252 have been met here. The Pentagon’s pretextual effort to disguise its punishment of Anthropic as national security enforcement is only highlighted by the fig leaf of administrative process with which it was issued. The statutory authorities to designate supply chain risks require a neutral assessment, interagency review, notice to the company, and notification to Congress. But here the required process never took place, and instead was replaced with a pontifical social media broadside targeting Anthropic’s corporate values and perceived ideology. Driven by Trump’s Truth Social condemnation of Anthropic as a “RADICAL LEFT WOKE COMPANY,” Hegseth justified the supply chain risk designation by claiming that “the Commander-in-Chief and the American people alone will determine the destiny of our armed forces, not unelected tech executives;” he labeled Anthropic’s refusals to tolerate mass surveillance or lethal autonomous killing as the “sanctimonious rhetoric of ‘effective altruism,’” or “cowardly act[s] of corporate virtue-signaling[.]” But if, as Hegseth claims, Anthropic’s AI program Claude presents such a security risk, why was it used the very next day (as the Wall Street Journal reported) in the Administration’s recent attack on Iran?
Anthropic’s Designation as Unconstitutional Punishment
Further, even if the statute could be read to authorize the Pentagon’s ad hoc punishment of Anthropic, it would be unconstitutional as applied. Like other punitive actions imposed by the Trump Administration, the statutory designation would both violate the First Amendment and work an unconstitutional exercise of a claimed power that our Constitution denies the President: the power to arbitrarily punish political enemies, which violates both constitutional rights and structure.
This rights violation is manifest, and has been spelled out in all the recent law firm decisions. The First Amendment bars the government from punishing Anthropic on the basis of speech and political values—as the Administration has blatantly done here, including by labeling Anthropic’s stated principles as so-called “Silicon Valley ideology.” Individuals and their companies have a right to believe and say what they believe, not just to parrot government orthodoxy. As the Supreme Court famously put it, “[i]f there is any fixed star in our constitutional constellation, it is that no official, high or petty, can prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion . . . .” As a private company in a free-market system engaged in contract negotiations with the federal government, Anthropic can seek to specify by contract how it will allow the Pentagon to deploy its AI products. The government may, of course, choose to reject those conditions and not contract on that basis. But the government cannot thereafter pretextually invoke national-security authorities to punish Anthropic for the views that inform its unwillingness to have its technology utilized in certain ways. Moreover, the government cannot seek to suppress Anthropic’s speech by coercing others not to contract with it because of its speech. In sum, Anthropic—like the many targeted law firms and universities that have all won their First Amendment lawsuits against the Trump Administration in federal court—has a cut-and-dry case that Hegseth’s stigmatizing and punitive supply chain risk designation amounts to unconstitutional retaliation for protected speech in violation of the First Amendment.
But beyond Anthropic’s First Amendment claim, the President’s attempt to target Anthropic also fails for another reason, embedded in the language and structure of the Constitution itself: the Constitution denies the president the unilateral power to punish his political enemies without fair trial. By targeting Anthropic—and others—Trump has asserted a power that, for centuries, has been denied to executives ranging from British kings to early colonial governors who sought to impose arbitrary punishment without the due process afforded by judicial trial. By targeting Anthropic for retaliation, the President claims the power to declare someone guilty and impose punishment, described at English common law and in the U.S. Constitution as the power to “attaint.” But even those kings and colonial governors recognized that they could not “attaint” someone without the passage of a “bill of attainder” by Parliament or the colonial legislature.
Prior posts have explained the backdrop of English monarchical practice and the use of bills of attainder during the American Revolution (including to confiscate Loyalist property). The Framers of the U.S. Constitution chose to outlaw bills of attainder entirely. By explicit constitutional language in two separate clauses, they denied the power to both the federal government and the states. But here, the President has sought to wield that forbidden power of attainder against Anthropic, as he has against other perceived enemies. Indeed, he has audaciously claimed power that is not only denied to him by the Constitution, but exceeds that exercised by English kings and their colonial governors, all of whom recognized that attainder required not just executive action, but legislative concurrence. If the President and Congress, acting together, cannot violate the Bill of Attainder Clause, then the President cannot constitutionally exercise the same power alone, as either inherent executive authority or as an authority allegedly delegated by Congress. Thus, no statute, regardless of whether it applies to the facts at hand—and not even ones centered on national security, such as the supply chain risk statutes invoked by Hegseth—can grant the Executive a power that Congress itself may not exercise.
Courts interpreting the Bill of Attainder Clause throughout American history have explained that a governmental action constitutes an unlawful attainder if it (1) applies with specificity and (2) imposes punishment without process. Secretary Hegseth’s designation of Anthropic as a supply chain risk, and his attempt to bar other private sector actors from engaging with Anthropic—without evidence or reason—easily meets both prongs. The designation not only singles out Anthropic, but punishes it: by extinguishing its right to compete for federal government contracts and by attempting to prevent it from doing business with any and all private customers who have contracts with the federal government.
Conclusion
In sum, the Pentagon’s decision to label Anthropic a supply chain risk illegally distorts § 3252 to mete out constitutionally forbidden punishment. If allowed to proceed, it would render the statute unconstitutional as applied. The executive’s obligation to protect U.S. national security does not include the right to impose arbitrary punishment upon those who reject his orthodoxy without judicial process. Statutory limits, First Amendment freedoms, and the Constitution’s absolute bar against attainder independently and jointly invalidate the Trump Administration’s latest salvo in its unprecedented, lawless campaign to use federal authority to punish perceived political enemies.
FEATURED IMAGE: A figurine in front of the logo of the U.S. artificial intelligence safety and research company Anthropic during a photo session in Paris on February 13, 2026. (Photo by Joel Saget / AFP via Getty Images)
Click Here For The Original Source
