Sarah Perez
Our Take:
Image and visual models are now the biggest download drivers for AI apps, far outpacing traditional conversational model updates; they create strong curiosity and trial installs even if they don’t reliably convert to paid users. The downloads-to-revenue disconnect matters — a big spike in installs often buys attention, not sustainable monetisation, with only a few players (notably OpenAI) successfully converting that interest into meaningful short-term spending.
Your Takeaway:
If you run or advise apps, treat image-model launches as high-impact user-acquisition plays that need a conversion plan baked in — optimize onboarding, trial-to-pay flows, and clear value props before release. Also, watch costs and privacy risks around image data and model access: don’t assume downloads justify long-term spend without a follow-through commercial and data-control strategy.
New image AI models are driving much more app growth than chatbot updates, with 6.5 times more downloads. Apps like ChatGPT and Google’s Gemini saw millions of new installs after releasing image models. However, more downloads don’t always mean more revenue, except in ChatGPT’s case.