On August 31, 2026, OpenAI released "A milestone in expanding access to AI" on its official website, revealing that ChatGPT Ads reached a $1 billion annualized revenue run rate in less than 200 days after launch, with tens of thousands of advertisers already using it. The advertising business has also expanded to more than 40 countries, and Ads Manager self-service purchases have been opened in India, Europe, the Middle East, and North Africa.
Here, "$1 billion" refers to annual revenue at current rates, not the company has already recognized $1 billion in advertising revenue. This standard better illustrates the growth rate and means ChatGPT's business model is starting from subscriptions and APIs, adding a stable revenue pillar for free and low-cost users.
Advertising is transitioning from a test slot to a complete delivery system
The early testing phase mainly answered "Do users accept ads appearing in conversations?" The previous internal ChatGPT ad test summary recorded the initial boundaries. Now, OpenAI's signals have changed: CPC and Result Optimization bids dominate most activities, Pixel, Conversions API, product catalogs, geographic targeting, and custom audiences have all become common capabilities, and small and medium-sized enterprises have become an important part of advertisers.
For advertisers, ChatGPT Ads are no longer just about buying an impression placement, but rather providing relevant business information during the decision-making process where users have already stated their goals, budget, and constraints. It may shorten the path from discovering needs to comparing products, but the results still rely on product data, conversion feedback, and audience settings, and cannot rely solely on ad copy.
The biggest variables remain the boundaries between independence and privacy
OpenAI emphasizes that ads are clearly labeled and separated from answers, advertisers cannot access private conversations, and ads do not change the content of their responses. The system can use the current conversational context; Whether to incorporate broader ChatGPT usage scenarios depends on the country and user settings.
This boundary directly affects two groups: free users must judge whether personalized settings meet expectations, and ad teams must prove that delivery relevance does not rely on reading private content. The faster the scale, the more it needs to continuously observe mismatches, sensitive scenario avoidance, and answer neutrality. For OpenAI, a $1 billion annualized run rate proves advertising can be profitable; long-term trust is what determines whether it can become a business on par with subscriptions.
