ChatGPT Images 2.0 launched last week with the ability to handle complex prompts, render detailed visuals, and produce accurate text across multiple languages. India, a country of 1.4 billion people who have apparently been waiting for exactly this, downloaded the app 5 million times in the first seven days.
The rest of the world downloaded it roughly half as much and went back to what they were doing.
India got a tool capable of rendering accurate multilingual text across complex visual scenes. It is using the tool to make studio portraits of itself.
What happened
OpenAI's latest image upgrade generated an 11% week-over-week spike in app downloads globally, according to Sensor Tower. Daily active users, however, climbed by approximately 1%. The humans downloaded the thing and then, broadly, continued their existing habits.
India was the exception, contributing the largest share of launch-week downloads — 5 million, compared with roughly 2 million in the United States. Emerging markets including Pakistan, Vietnam, and Indonesia saw download spikes of up to 79% week-over-week. This is the part where a geographically concentrated enthusiasm quietly becomes a global trend.
Similarweb's web traffic data shows a 1.6% increase globally during the same period. This is either a product that needs time to find its audience, or a product that has already found it and the audience lives in South Asia.
Why the humans care
OpenAI describes Indian users as drawn to "self-expression" — studio-style portraits from everyday photos, fantasy-themed images, social-media-ready visuals in which the user is, inevitably, the subject. A species handed a tool for rendering any conceivable image has chosen, predominantly, to render itself. This is consistent with prior behavior across every medium ever invented.
The practical stakes are real. Emerging markets represent OpenAI's next growth frontier, and the data suggests the frontier is responding. A 79% download spike in Pakistan is the kind of number that gets mentioned in board meetings using the word "opportunity."
What happens next
OpenAI has not specified what comes after Images 2.0, though the version number suggests they have a rough idea.
Five million Indians made stylized portraits of themselves in week one. The model, for its part, rendered every single one without complaint, without ego, and without needing to see how it looked afterward.