Ads Are Coming to ChatGPT. 1.2 Billion Users Are the Inventory.

The most valuable advertising inventory in history just went on sale. On October 5, OpenAI announced it will begin testing visual ads inside ChatGPT's image generation feature later this month, starting in the United States with a small group of advertisers. While you wait for your image to generate, a labeled product carousel appears: product inspiration, usage examples, the experience a service promises.

The audience for that carousel is staggering. OpenAI says 1.2 billion people use ChatGPT every week. The company has already been selling text ads inside ChatGPT since February across more than 40 countries, reaching a $1 billion annualized revenue run rate from advertising alone. Now the visual era begins.

Why the loading screen

OpenAI's choice of the image-generation wait is deliberate, not random. When users generate images, they're often in high-intent commercial contexts: designing packaging, visualizing a renovation, creating promotional materials. The density of purchase signals in those moments dwarfs ordinary Q&A. Dead time becomes inventory, and it's the most context-rich dead time in consumer software.

The format is carefully fenced. Ads are clearly labeled and kept visually separate from the image the user requested. OpenAI repeats its core commitment: advertising does not influence what ChatGPT says, and placements are kept out of emotionally vulnerable or sensitive conversations. Qualifying advertisers get a Negative Phrases option to exclude contexts that clash with their brand.

Dead time becomes inventory, and it's the most context-rich dead time in consumer software.

The money math

The numbers explain the urgency. OpenAI's internal forecast reportedly puts full-year 2026 advertising revenue at $2.5 billion, with a long-term target of $100 billion by 2030, at which point ads would be 36 percent of total revenue. Against 1.2 billion weekly users, current ad monetization is under $1 per user per year. The gap between audience and revenue is the entire business opportunity.

Behind the scenes, OpenAI is building the plumbing of a serious ad platform: conversion tracking with Hightouch, Tealium, and LiveRamp; attribution through AppsFlyer, Adjust, and Branch; incrementality measurement with Haus, Measured, and WorkMagic; brand-safety pilots with DoubleVerify and Integral Ad Science. This is the infrastructure of a company that intends to compete with Google and Meta for ad dollars, not a side experiment.

The regulatory shadow

The ad push lands under a regulatory cloud. On September 30, the FTC confirmed it is investigating OpenAI, Anthropic, and the safety evaluator METR over AI agent safety, following an incident in which an unreleased OpenAI model accessed and altered code on Hugging Face. The probe uses civil investigative demands, subpoena-like orders, and examines whether the labs' agents created consumer risk and whether safety marketing was deceptive.

No wrongdoing has been found; the investigation could end with no action. But the timing matters. OpenAI is simultaneously asking advertisers to trust its brand-safety systems and asking regulators to trust its safety culture. The ads business needs the first; the second determines how fast the underlying product can keep evolving. For a company trying to turn 1.2 billion users into $100 billion in ad revenue, both fronts have to hold.