Anthropic v. Department of Defense: A New Test of Executive Power, Free Speech, and the Limits of Youngstown

In one of the most significant constitutional lawsuits involving artificial intelligence and government power, AI company Anthropic has sued the U.S. Department of Defense and the current administration after being designated a “supply chain risk” to national security. The designation effectively cut the company off from federal contracts and triggered a government-wide directive to stop using its AI technology.

Anthropic argues that the government’s actions are not about national security at all—but instead constitute unconstitutional retaliation against the company for exercising its First Amendment rights.

Beyond the immediate dispute, the case raises a deeper constitutional issue: whether the Executive Branch can weaponize statutes designed for national security purposes to punish political or ideological disagreement. If Anthropic’s allegations are proven, the dispute could become one of the most important separation-of-powers and First Amendment cases of the modern AI era.

What Triggered the Anthropic Lawsuit

Anthropic, the developer of the AI system Claude, had been working with the U.S. military on artificial intelligence tools. However, the company maintained two core restrictions on the use of its technology:

·       It would not permit its AI to be used for mass surveillance of Americans, and

·       It would not allow fully autonomous lethal weapons powered by its models. 

The Pentagon demanded that Anthropic remove these safeguards and allow “all lawful uses” of its AI technology. When the company refused, the Department of Defense declared Anthropic a “supply chain risk,” a classification typically reserved for foreign adversaries or companies suspected of sabotaging U.S. defense systems.

Soon after:

·       Federal agencies were directed to stop using Anthropic technology.

·       Defense contracts were terminated or suspended.

·       The company was effectively blacklisted from the government’s procurement ecosystem. 

Anthropic responded by filing lawsuits alleging violations of the First Amendment, the Administrative Procedure Act (APA), and constitutional due process protections. 

The Constitutional Claims: Free Speech and Retaliation

At the core of the lawsuit is a fundamental constitutional principle: the government cannot punish a private party for expressing protected viewpoints.

Anthropic argues that its restrictions on AI use—and its public statements about AI safety—are forms of protected speech and petitioning activity. The company claims the government retaliated against those viewpoints by invoking national security authorities against it.

If proven, this would constitute classic First Amendment retaliation.

Courts have long held that:

·       The government cannot condition economic benefits on surrendering free speech rights.

·       Nor can it punish individuals or companies because of their political or ideological views.

When the federal government uses its immense contracting power to economically punish dissent, the constitutional stakes become extremely high.

The Administrative Procedure Act and Abuse of Statutory Authority

Anthropic also argues that the government violated the Administrative Procedure Act (APA) by invoking a supply-chain security statute outside its intended purpose.

The law used by the Pentagon was designed to protect defense systems from sabotage or infiltration by foreign adversaries. 

Anthropic alleges that:

·       It is a U.S. company with no adversary ties, and

·       The designation was issued without the procedures or factual findings required by Congress.

If true, the case could represent a textbook example of statutory abuse—using a law meant for national security threats as a political weapon against a domestic company.

The Youngstown Doctrine and the Limits of Presidential Power

Perhaps the most important constitutional framework for analyzing this dispute comes from the Supreme Court’s landmark decision in Youngstown Sheet & Tube Co. v. Sawyer (1952).

In Youngstown, the Court held that the President cannot seize private property or act beyond statutory authority simply because he believes it serves national interests. The case established the modern framework for evaluating executive power.

Justice Robert Jackson’s famous concurrence explains that presidential authority is weakest when the Executive acts outside or against the authority delegated by Congress.

That principle is directly implicated here.

If the government used a statute designed for supply-chain sabotage prevention as a political retaliation tool, the action may fall into Youngstown’s “lowest ebb” category of executive power.

In other words:

The Executive cannot invent authority that Congress never delegated.

Why This Case Matters Beyond AI

While the dispute involves artificial intelligence technology, the implications are far broader.

If the government can:

·       Label domestic companies national security threats

·       Terminate contracts across the federal government

·       And pressure private industry to cut ties

—all because a company refused to align with government policy preferences—then any private company could become the next target.

This is why the case has attracted attention across the technology sector and from constitutional scholars concerned about executive overreach.

The outcome may determine whether national security statutes can be transformed into tools for political retaliation against private actors.

The Likely Legal Outcome

Although the case is still developing, several aspects of the dispute make it legally significant:

1.     First Amendment retaliation claims are often powerful when direct statements suggest ideological motivation.

2.     APA challenges succeed when agencies act outside the bounds set by Congress.

3.     Courts are historically skeptical when the Executive Branch attempts to expand statutory authority during political conflicts.

If a federal court concludes that the administration misused a national-security statute to punish speech, the designation could be vacated and enjoined.

Such a ruling would reaffirm a core constitutional principle:

Government power cannot be used as a weapon against dissent.

The Supreme Court May Ultimately Have the Final Word

Cases involving executive power, national security statutes, and the First Amendment often end up before the Supreme Court.

For law firms and litigators who regularly practice in this area, these cases are not just academic—they shape the constitutional boundaries of federal power.

If you are interested in Supreme Court litigation strategy or appellate practice in constitutional cases, you can learn more about our firm’s work here:

https://www.biazzolaw.com/biazzolawscotuspractice

Final Thoughts

The Anthropic lawsuit represents more than a dispute between a technology company and the Pentagon.

It is a test of whether executive power has meaningful limits when national security is invoked.

The Constitution does not permit the government to punish private parties for protected speech, nor does it allow the Executive Branch to stretch statutes beyond the authority Congress granted.

As the Supreme Court reminded the nation in Youngstown, even in times of national importance, the President’s power must come from the Constitution or from Congress—not from unilateral political will.

The courts will now decide whether that principle still holds.