What Happened
App uninstall tracking data shows ChatGPT's uninstall rate jumped 295% in the period following news of OpenAI's agreement with the Department of Defense, according to TechCrunch. During the same window, Claude's downloads rose. The figures reflect a measurable consumer reaction to a policy controversy that broke in late February 2026, involving OpenAI's willingness to accept broader military use terms that Anthropic had refused.
Why It Matters
Uninstall spikes following PR crises are common but rarely reach this scale. A 295% increase suggests the DoD news reached a broad consumer audience, not just the tech-native users who monitor policy developments closely. The timing correlation with Claude's download increase implies that some users made an active choice to switch products rather than simply abandoning AI chatbots altogether.
This matters for the competitive dynamics between ChatGPT and Claude. Both tools have broadly similar capabilities for everyday tasks, meaning user stickiness depends heavily on brand trust and perceived values alignment. OpenAI's deal with the Pentagon created an opening for Anthropic, which was simultaneously framing itself as the company that held firm on military use restrictions at real commercial cost.
The scale of the uninstall spike also tells something about the composition of ChatGPT's install base. Core power users are less likely to uninstall over a policy story - they've built habits and workflows around the tool. A 295% spike points to a large segment of casual users who respond to news cycles and don't have deep switching costs.
The simultaneous rise in Claude downloads is the more commercially significant data point. Users didn't just leave; a meaningful portion moved to a specific competitor. That's an active switching decision, not just passive churn.
Our Take
App uninstall numbers should be read carefully - a spike can revert quickly as news cycles move on. But the direction is real: OpenAI's DoD deal caused measurable brand damage with consumers in the short term. Anthropic's positioning as the company that refused military surveillance use cases - even at the cost of a federal supply-chain designation - appears to have paid off in user perception. Whether Claude retains those new users depends on the product experience, not the politics. Retention data from 30 and 90 days out will tell the real story.
For product teams at both companies, the data is actionable in different ways. OpenAI needs to understand whether the churned users were high-value or casual, and whether they return as the news cycle moves on. Anthropic needs to measure session depth and return frequency among the new signups to determine whether the migration is sticky.