On March 1, 2026, something unprecedented happened: Claude overtook ChatGPT as the #1 app in the US App Store. Claude reached #1 in 16 countries. Daily active users surged from 4 million in January to 11.3 million in early March — a 183% increase in two months. The trigger was not a product launch or a marketing campaign. It was an ethical stand on autonomous weapons that turned into one of the most-discussed stories in AI in 2026.
What Happened: The Pentagon Deal and Anthropic's Response
The story begins with the US Department of Defense seeking AI partnerships for classified military applications. OpenAI agreed to a deal that included loosened restrictions on the use of its models for defense applications. Anthropic refused a similar deal: Anthropic drew three red lines — it would not allow Claude to be used for mass surveillance, autonomous lethal weapon systems, or applications that removed core ethical safeguards from the models.
The Pentagon's response was to list Anthropic as a 'supply chain risk' and cancel its contract with the company. Hours later, OpenAI signed its own Pentagon deal. CEO Dario Amodei publicly stated that Anthropic does not believe the Pentagon's designation is 'legally sound' and vowed to challenge it. The contrast was clear: one company drew ethical lines and lost a government contract; another accepted the deal.
The QuitGPT Boycott: What Happened to ChatGPT
- Over 2.5 million people pledged to cancel their ChatGPT subscriptions following news of OpenAI's Pentagon deal.
- ChatGPT app uninstallation rates jumped 295% over the weekend following the announcement.
- Claude hit #1 in US App Store downloads — the first time it had ever surpassed ChatGPT on that metric.
- Even within OpenAI, public dissent emerged: research scientist Aiden McLaughlin wrote on X: 'I personally don't think this deal was worth it.'
- Singer Katy Perry publicly mentioned switching to Claude — a signal of the story's mainstream cultural reach.
What Anthropic Did With the Influx of New Users
A 183% surge in daily active users in two months creates a specific problem: new users hit usage limits quickly, feel frustrated, and leave. Anthropic responded with two moves. First, it released Claude's Projects and Artifacts features to free users — turning Claude from a chatbot into a workspace. Second, from March 13 to March 27, 2026, Anthropic doubled usage limits for all Free, Pro, Max, and Team users during off-peak hours (outside 8 AM–2 PM Eastern on weekdays, and all weekend). The extra usage does not count toward weekly limits.
What This Means for AI Users: Practical Takeaways
| Question | Short Answer | Detail |
|---|---|---|
| Is Claude better than ChatGPT now? | Different strengths | Claude Sonnet 4.6 leads SWE-bench (76.8%); GPT-5.4 leads professional knowledge work (83% GDPval). Neither wins overall. |
| Should I delete ChatGPT? | Personal decision | The Pentagon deal is a real ethical concern. Functionally both remain excellent. Many users now use both for complementary tasks. |
| Is Claude now free to use without limits? | No | The 2x promo runs until March 27 only. After that, free tier returns to 5 messages/day. Paid tiers have rolling five-hour usage limits. |
| Did OpenAI do anything wrong? | Contested | OpenAI's deal includes restrictions (no autonomous weapons, no mass surveillance). Where lines should be drawn is a legitimate ongoing debate. |
Pro Tip: If you have never tried Claude, this week is the best time — the 2x usage promo ends March 27. Create a free account at claude.ai and select Claude Sonnet 4.6. The most revealing test: paste a 5,000-word document and ask it to analyze the argument structure. This is where Claude's long-context reasoning stands out most clearly against other models.
Claude's Growth in India: What the Numbers Show
Claude's surge was not limited to the US. The App Store #1 position included 16 countries, and India has shown particularly strong adoption trends among students and developers. Claude Sonnet 4.6's combination of strong coding ability (SWE-bench 76.8%), nuanced writing quality, and the 1-million-token context window has made it the model of choice for Indian B.Tech students building final-year projects, researchers processing academic literature, and developers at Indian AI startups who need a reliable reasoning model for complex agentic tasks.