ChatGPT's download growth has decelerated sharply. According to Sensor Tower data, ChatGPT app downloads grew just 14% year-on-year in April 2026. In the same period, uninstalls jumped 132% year-on-year. March was worse: uninstalls were up 413% compared to March 2025, a spike partly attributed to a Pentagon partnership announcement that alienated a portion of the user base.
Monthly active user growth tells a similar story. Growth moderated from 168% year-on-year in January to 78% in April. Still positive, but the trajectory is pointing the wrong way ahead of a planned OpenAI IPO.
The competitive context makes the numbers harder to dismiss. Anthropic's Claude recorded download growth of approximately 1,000% year-on-year in the same period. OpenAI is not losing to a theoretical competitor. It is losing downloads to a product that is actively gaining them.
OpenAI recently missed internal targets for new users and revenue, prompting reported concerns from CFO Sarah Friar about future compute costs. At the same time, the Musk v. Altman trial is actively running in court, with documents and emails being revealed that shed light on OpenAI's internal history and strategic direction.
Year-on-year spike in ChatGPT uninstalls in March 2026, following a Pentagon partnership announcement
Why it matters
ChatGPT's deceleration matters because it complicates the dominant narrative around AI adoption: that growth is linear, compounding and inevitable. The data shows it is none of those things for individual products.
For Australian marketers using AI tools, the competitive shift has practical implications. Claude's 1,000% download growth is not driven by marketing. It is driven by product quality, pricing decisions and the Anthropic API ecosystem that has made Claude the default choice in a growing number of enterprise and developer workflows. If your team defaulted to ChatGPT because it was first, the current landscape warrants a reassessment.
The broader lesson is about platform dependency. Any business that has built workflows around a single AI tool is carrying concentration risk. The market is moving fast enough that the market leader in January may not be the market leader in January 2027.
OpenAI still leads on monthly active users and consumer brand recognition. But the gap is narrowing, and the uninstall data suggests active dissatisfaction, not just slower new user acquisition.
What to do about it
The AI product market is competitive in a way that the first two years of ChatGPT's dominance obscured. The April 2026 data is a correction to that story.
