Anthropic got banned by the government and hit #1 in the App Store

In four days, Anthropic went from a company the Pentagon was trying to bully into submission to the most downloaded app in the United States. The Trump administration banned them. Trump called them "radical left nut jobs" on Truth Social. Hegseth labeled them a supply-chain risk to national security. OpenAI swooped in to take the contract they walked away from.

And then millions of people downloaded Claude.

This is the most consequential AI news week of 2026. Let's dig in.

⚡ The Big Story: Anthropic Drew a Line

Anthropic had been negotiating a Department of Defense contract worth up to $200 million. They tried to get two carve-outs written in: no use of Claude for mass domestic surveillance of Americans, and no use of Claude to power fully autonomous weapons systems. The Pentagon said no. Anthropic said then we can't sign.

Defense Secretary Pete Hegseth responded by threatening to invoke the Defense Production Act, which would have forced Anthropic to hand over their models. Anthropic held the line. Hegseth designated them a supply-chain risk to national security. Trump called them "radical left nut jobs" and directed every federal agency to immediately stop using their technology.

Hours later, OpenAI announced they had signed their own Pentagon deal. Sam Altman posted that OpenAI's contract included the same red lines Anthropic had requested. The backlash was immediate - users flooded Reddit and X with guides for canceling ChatGPT. Katy Perry posted Claude's Pro subscription page with a heart emoji.

By Saturday morning, Claude was #1 in the App Store. ChatGPT dropped to second. Anthropic said free users have grown 60%+ since January and daily sign-ups tripled since November.

My take: Two things happening at once. First, the tactical question: was OpenAI's deal actually different from what Anthropic refused? Probably not much. OpenAI's contract relies on existing laws rather than explicit contractual prohibitions. Altman admitted the deal was "definitely rushed" and "the optics don't look good."

Second, the strategic question: Anthropic lost a $200M contract. They also got millions of new users and a massive PR moment at exactly the time they're planning an IPO. Amazon's $8 billion bet is now worth roughly $13.8 billion on paper. Bank of America analysts project Anthropic could add $6 billion in incremental AWS revenue in 2026.

Dario Amodei built a company around the thesis that safety and capability aren't in conflict. On Friday, that thesis converted into App Store rank #1.

🎮 NVIDIA's Coming-Out Party: GTC 2026

Jensen Huang keynotes from SAP Center in San Jose on March 16. 30,000 attendees from 190 countries.

NVIDIA's Q4 FY2026 revenue: $68.1 billion, up 73% YoY. Huang reportedly will "unveil a chip that will surprise the world" at GTC. NVIDIA's next data center platform is Vera Rubin (HBM4 memory), but the rumored inference-focused chip would be something new.

The next wave is inference - the compute required to run models at scale for billions of users. That market is an order of magnitude larger than training.

🎯 Quick Hits

• Anti-AI protesters marched through London's King's Cross tech hub - Organized by Pause AI and Pull the Plug, past UK headquarters of OpenAI, Meta, and Google DeepMind
• Apple's reimagined Siri is powered by Google Gemini - Gemini's 1.2-trillion-parameter model for new context-aware Siri, running on Apple's Private Cloud Compute. Target launch: March 2026 with iOS 26.4.
• Claude Opus 4.6 is top-ranked model on Artificial Analysis benchmarks - Beating GPT-5.2 and Gemini. Also #1 on Arena.ai.
• OpenAI has 900 million weekly active users - Roughly 11% of every person alive
• Samsung targeting 800 million Gemini-equipped devices by end of 2026

💭 One Thing I'm Thinking About

This week proved that an AI company's values are now a product feature, whether they want them to be or not. Anthropic held a position under political pressure, got punished for it, and users responded by making them the most downloaded app in the country. The $200M contract loss turned into a marketing event worth multiples of that in brand value.

Trust is a competitive advantage, and it accumulates slowly until it pays off all at once.

The uncomfortable part: Sam Altman told his own employees he shared Anthropic's red lines. Then he signed a deal without those red lines explicitly protected. We're entering a phase where AI companies have to decide who they are, not just what they build.