The app that made "AI" a household word is watching users walk away faster than new ones sign up.
The Summary
- ChatGPT's uninstall rate jumped 132% year-over-year in April, with March seeing a 413% spike following OpenAI's Pentagon deal
- Monthly active user growth collapsed from 168% in January to 78% in April, per Sensor Tower data
- The slowdown arrives just as OpenAI eyes an IPO, raising hard questions about whether consumer AI apps can sustain venture-scale growth
The Signal
ChatGPT's uninstall problem isn't just about one app losing steam. It's the first real stress test of whether AI chatbots can hold consumer attention once the novelty fades. The March spike, tied directly to the Pentagon announcement, tells you something important: people still see these tools as products they can opt out of, not infrastructure they depend on.
Compare that to how quickly email, search, or social media became mandatory. Nobody uninstalls Gmail at 400% year-over-year rates because they disagree with Google's government contracts. The uninstall spike suggests ChatGPT hasn't crossed the moat from "interesting tool" to "can't live without it."
"ChatGPT still has a substantially larger user base than rivals, but growth rate is the metric that matters for IPO pricing."
The timing couldn't be worse for OpenAI's public market ambitions. IPOs live and die on growth narratives. Investors price future revenue based on user acquisition trends, not absolute numbers. A company adding users at 78% year-over-year sounds healthy until you realize it was 168% three months ago. That's deceleration, and public markets punish deceleration harder than they reward absolute scale.
Here's what the data pattern reveals:
- January: 168% MAU growth (pre-controversy baseline)
- February: Pentagon deal announced
- March: 413% uninstall spike (user backlash)
- April: 78% MAU growth, 132% uninstalls (sustained churn)
The February Pentagon deal wasn't just a PR problem. It was a market research accident that revealed ChatGPT's fragile product-market fit. When a non-product decision causes triple-digit uninstall spikes, you don't have infrastructure. You have a discretionary app competing for attention with every other thing on someone's phone.
OpenAI's real problem isn't competition from other chatbots. It's that chatbots themselves might not be sticky enough products to justify the valuations everyone's been throwing around. The agent economy needs agents that do things, not chatbots you consult. ChatGPT's slowdown suggests users figured that out before the market did.
The Implication
If you're building in AI, watch what happens with OpenAI's IPO timing. A delayed or down-round IPO will ripple through every AI company's valuation model. The message: consumer AI apps need to become infrastructure fast, or they'll get valued like media companies, not platform companies.
For users, the uninstall trend is actually clarifying. The tools worth keeping are the ones you forget are AI. The ones that do something without asking you to think about prompts or check answers. ChatGPT's decline is the market saying: stop making me work for it.