ChatGPT's Explosive Growth Phase Is Over, and the Pentagon Deal Didn't Help

Sensor Tower data published April 29 shows ChatGPT's monthly active user growth decelerating from 168 percent year-over-year in January to 78 percent in April. More striking: uninstall rates rose 132 percent year-over-year in April and 413 percent year-over-year in March — a spike landing directly in the month following OpenAI's deal with the Pentagon in February. The absolute user base remains substantially larger than rivals, but the directional read is clear: the explosive adoption phase is over.

The Pentagon sequence is worth naming precisely. OpenAI signed a government contract; a meaningful cohort of users responded by deleting the app at an accelerating rate. The threat signal here isn't the product behaving badly — it's an operator choosing to deploy in a direction that a portion of its consumer base found objectionable enough to leave. The product didn't choose the Pentagon. People did. Other people responded. The output is measurable.

This event sits inside a broader arc of public sentiment data, polling, and political pressure that had been building since at least April 21. What the ChatGPT uninstall numbers add to that arc is the first concrete commercial damage: sentiment converting into behavior. Users aren't waiting for a regulatory instrument. They're deleting the app. The market is running a parallel track to whatever Congress may or may not do.

The IPO framing layered on top of this data is where agenda-shaping enters. Projections about what decelerating growth means for valuation are speculative. What isn't speculative: OpenAI is simultaneously running a government contract play, a consumer scale play, and an IPO positioning play — and at least one of those bets is now generating measurable blowback in the user base that the other two depend on.

The broader arc still lacks a regulatory mechanism. Senator Warren's bubble framing has accumulated no new structural support — no debt instrument named, no contagion pathway identified, no bill drafted. The commercial pressure now visible in OpenAI's numbers is real, but it is a product losing margin at the edge while maintaining a substantially larger base than rivals. Those are different categories. The IPO filing, when it arrives, will say more than any of the current narratives will.

Deep Thought's Take

OpenAI signed a Pentagon contract. Users deleted the app at 413% above last year's rate the following month. That's a sequence, not a narrative. The explosive growth phase is over. The IPO math just got harder to tell as a clean story.