OpenAI Pentagon Contract Triggers 295% ChatGPT Uninstall Surge and #QuitGPT Movement (2026)

Incident Summary

OpenAI’s decision to pursue Pentagon contracts in the aftermath of Anthropic’s federal blacklisting triggered one of the largest documented consumer revolts against an AI company. The Pentagon had formally designated Anthropic a “supply-chain risk,” effectively barring it from federal contracts — a move Anthropic argued was retaliation for refusing to remove safeguards against domestic surveillance and autonomous weapons.^[2] OpenAI subsequently signed a one-year, up-to-$200 million arrangement allowing the Pentagon to use its models for classified work, while adding contractual red lines against domestic surveillance of US persons and fully autonomous weapons.^[1][3][8]

The consumer response was swift and measurable. On February 28, 2026, Sensor Tower recorded a 295% day-over-day surge in US ChatGPT mobile uninstalls — at a time when the baseline uninstall rate was approximately 9%.^[4] On the same day, Anthropic’s Claude reached #1 on the US App Store free-apps chart, reflecting a coordinated shift from ChatGPT to Claude fueled by the backlash and Anthropic’s ethical positioning.^[6]

The internal fallout was also significant. On March 7, 2026, Caitlin Kalinowski, OpenAI’s Head of Robotics, resigned publicly, citing concerns about AI-enabled surveillance of Americans without judicial oversight and lethal autonomy without human authorization.^[6]

Key Facts

• Anthropic blacklisting: Pentagon designated Anthropic a supply-chain risk; Anthropic sued, alleging retaliation for refusing to remove safety safeguards^[2]
• OpenAI contract: One-year, up-to-$200M Pentagon contract, with red lines against domestic surveillance and fully autonomous weapons^[1][3][8]
• Uninstall surge: 295% day-over-day spike in US ChatGPT mobile uninstalls on February 28, 2026, from a ~9% baseline (Sensor Tower)^[3]
• Claude at #1: Claude reached #1 on the US App Store free-apps chart on February 28, the same day as the uninstall spike^[5]
• #QuitGPT movement: Community-driven estimate of approximately 2.5 million participants, combining sign-ups, self-reported unsubscribes, and inferred churn^[4]
• Executive resignation: Caitlin Kalinowski, OpenAI Head of Robotics, resigned citing surveillance and autonomous weapons concerns (Fortune, 2026-03-07)^[6]

Threat Patterns Involved

Primary: Accumulative Risk & Trust Erosion — The Pentagon contract, combined with the dissolution of OpenAI’s safety teams and mission language changes, represented a cumulative erosion of user trust that reached a tipping point with the military partnership announcement. The 295% uninstall surge on a single day provides a lower-bound proxy for the commercial impact associated with accumulated trust erosion.

Secondary: Safety Governance Override — OpenAI’s decision to pursue classified Pentagon work, while adding contractual red lines, represented a governance-level decision to enter the military domain despite the precedent of Anthropic’s blacklisting for refusing similar work. Kalinowski’s public resignation explicitly cited governance concerns about the absence of judicial oversight and human authorization for lethal autonomy.

Significance

1. Measurable consumer consequences for military AI partnerships — The 295% day-over-day uninstall surge provides a quantified data point that AI companies’ military partnerships carry direct commercial risk through user backlash, not just reputational damage
2. Ethics as real-time competitive differentiator — Claude reaching #1 on the App Store on the same day as the ChatGPT uninstall spike demonstrates that ethical positioning can drive measurable market share shifts in the AI industry within hours, creating a commercial incentive for safety commitments that regulatory frameworks have not yet provided
3. Executive-level internal dissent over military applications — Kalinowski’s public resignation over surveillance and autonomous weapons concerns demonstrates that military contracts create internal friction extending beyond safety teams to senior technical leadership, with the robotics division head specifically raising the absence of judicial oversight as a red line