The Pentagon is turning procurement into policy, staking its claim as Washington's most powerful AI player with its decision to ditch Anthropic.

Why it matters: As lawmakers debate guardrails, the Pentagon showed how it can reshape the AI industry with a single contract decision.

• The Defense Department is the federal government's biggest tech buyer, and its requirements for companies to win contracts can become de facto rules and spread beyond military systems.
• In a regulatory vacuum, defense contracts speak the loudest.

What they're saying: "The biggest question is: What kind of business partner does the government want to be?" Jessica Tillipman, associate dean for government procurement law studies at George Washington University, told Ashley.

• "They need the AI companies," she said. "The government's a superpower ... but here it's trying to jam a lot of policy."

The big picture: The Trump administration has touted an anti-regulation, pro-AI innovation stance. But it has still regulated AI, just in a different way, former Office of Science and Technology Policy chief Alondra Nelson writes in Science:

• That includes, per Nelson, "intensive state intervention operating through industrial policy, trade restrictions, immigration controls, equity stakes in private firms (selected by the state), the redirection of research funding, and the strategic preemption of state authority."

Zoom in: The Pentagon's power play is rare, rests on questionable legal ground and could have regulatory implications beyond one company.

• Its designation of Anthropic as a "supply chain risk," which means companies must stop using Claude in cases directly tied to the department, is usually reserved for foreign adversaries. Anthropic is suing on the grounds that the Pentagon is violating its free speech rights and lacks congressional authority.
• It's also unclear how regulation-by-contract fits with the administration's AI action plan, which prioritizes rapid AI development and an industry-friendly approach.
• OSTP did not respond to a request for comment.

The Pentagon's actions are hitting Anthropic's contracts far beyond the government.

• At least 100 customers, from pharma to fintech companies, have asked to pause or cancel their contracts, Anthropic lawyer Michael Mongan said during a hearing this week.
• Microsoft is asking the court for a temporary restraining order. Otherwise, Microsoft argues, tech companies will have to act "immediately" to alter their products and contracts, which could potentially "hamper" soldiers.
• A hearing on whether to grant Anthropic temporary relief is set for March 24.

What we're watching: Regulation-by-contract will continue as AI companies seek out opportunities with the government, especially if new draft guidance from the General Services Administration adding "all lawful uses" language to procurement guidelines holds.

The bottom line: The Pentagon's move risks undercutting the White House's declared hands-off, pro-industry approach meant to supercharge AI growth.

• It could also create a new contract-by-contract approach to how AI is governed and controlled, leaving AI companies in the dark about how to work with the government.

The Justice Department's settlement with Ticketmaster this week marks a new chapter of the MAGA antitrust movement that may look different than some were expecting.

Why it matters: The deal offers a window into how Trump-era antitrust works, signaling that aggressive corporate breakups are out of favor.

• That could shape how the government approaches other cases — including in AI and emerging tech.

What they're saying: "The big thing is: We're not the Biden administration," a senior DOJ official told Ashley. "We're willing to bring cases. But we're also not afraid to settle. I don't think it's a secret to say this administration wants deals."

• The official called the proposed settlement with Ticketmaster "very meaningful," and said those describing it as weak would have only accepted a breakup, which a federal judge has not ordered in decades.

Driving the news: Just days into the trial, the Justice Department reached a surprise settlement with Ticketmaster. The terms have not yet been finalized, but some experts say the proposal to cap fees and open up Ticketmaster's platform to rivals is weak and disappointing.

• That's notable in a case with massive bipartisan interest fueled by public anger over "bots," ticket scalpers and the struggle to secure affordable tickets to concerts, sports and other live events.
• The settlement comes just weeks after former DOJ antitrust chief Gail Slater left the department. The senior DOJ official said Slater supported a settlement that did not include a breakup of Ticketmaster prior to her departure.
• Dozens of state attorneys general are still pursuing their own case against Ticketmaster.

One former DOJ official told Ashley that breaking up Ticketmaster would have been a "political layup."

What we're watching: Other companies looking to consolidate, buy up rivals or merge — including in AI — may be seeing a green light from Washington.

• "A little bit of modest enforcement could have gone a long way to preventing the kind of concentration that we see today," the former DOJ official said.
• "The same is true in AI where a little bit of enforcement up front could save a lot of difficult litigation later on."

Keep reading.

The Trump administration has ambitious deadlines to accelerate scientific discovery, Energy Department undersecretary Darío Gil and Dell CEO Michael Dell told Maria in an interview.

Why it matters: Government agencies and companies are mobilizing to use AI for energy, drug discovery and national security, all while promising new jobs.

Driving the news: The Genesis Mission, unveiled late last year, is a federal initiative led by the DOE and its national labs to use AI and emerging tech to accelerate scientific discovery.

• Gil said that includes quantum computers that aren't error prone by 2028, commercially viable fusion power plants in the 2030s, and a trained workforce of 100,000 scientists and engineers within the next decade.
• The goal is to double the productivity and impact of science R&D spending, he added.
• That means building on AI's success with language and code, and advancing its capabilities in mathematics, physics, chemistry, biology, engineering and more.

Read more.

Here's our guide to catch you up on the AI policy news you may have missed this week:

💼 Bipartisan AI push: Sens. Mark Warner (D-Va.) and Mike Rounds (R-S.D.) introduced new legislation focused on AI and the workforce, Ashley reported.

• The Economy of the Future Commission Act would create a panel to develop policy recommendations on AI's economic and workforce impacts.

🇨🇳 China's AI approach: Brookings' Kyle Chan wrote about China's "fundamentally different bet on how AI will shape the future."

• Instead of focusing on the race to AGI, "Chinese AI developers are racing along other axes of progress: efficiency, adoption, and physical integration, driven by both industry constraints and Beijing's policy focus."

⏰ Missed AI deadline: President Trump's state AI executive order directed his administration to announce this week which state-level AI laws it identified as "onerous" and a target for litigation — but that deadline passed without any action.

• The Commerce Department did not respond to requests for comment.