On October 21, 2026, OpenAI simultaneously connected 26 new countries to the ChatGPT Ads advertising platform, expanding its reach to 89 markets—from Argentina, Chile, and Colombia to Ukraine, Nigeria, and Pakistan. In the eight months since the launch of the US pilot, advertising within ChatGPT has evolved into a segment with a $1 billion annual run-rate, but ads remain visible only to Free and Go plan users, while Plus, Pro, and Enterprise subscribers remain ad-free.

image

What happened

On October 21, 2026, OpenAI connected 26 new countries to ChatGPT Ads at once. Among them are 14 Latin American countries, including Argentina, Chile, Colombia, and Peru, as well as Nigeria, South Africa, Serbia, Ukraine, Cambodia, Bangladesh, and Pakistan. Since the first tests in the US in February 2026, over 100,000 advertisers have launched campaigns on the platform, although there were about 20,000 in late August: a growth of approximately five times in two months. As of early October, approximately 60,000 remained active—an estimate The Information cites from Dan Salmon of New Street Research. The annual run-rate of OpenAI's advertising segment reached $1 billion in August, compared to $100 million in March 2026.

Context

The difference between ChatGPT Ads and traditional internet advertising is that it monetizes not link results, but the "purchase" dialogues with the assistant itself: OpenAI sells advertisers access to moments when a user is already formulating a product request. The scheme repeats the classic model of free services—ads subsidize free access, and the absence of ads becomes an argument for paid subscriptions—directly positioning OpenAI against the advertising markets of Google and Meta. The scale of its ambitions is also reflected in an unmet goal: the $2.4 billion in advertising revenue announced for 2026 has not yet been achieved by the platform. The key unfinished part is measurement: attribution and ad ranking are developing in parallel with expansion, and their maturity will determine whether ChatGPT becomes a full-fledged channel on par with search and social networks.

Why this matters for the industry

Advertisers are getting a new type of inventory—ad slots within dialogues, which are not available to search engines or social networks. For agencies and builders, a window for early testing has opened: competition for impressions in the 26 new markets is minimal, so hypotheses in Latin America, Africa, and South Asia can be tested more cheaply than in the US. A product perimeter is forming around the platform: creative generation for new formats, campaign automation, and analytics of sponsored brand presence versus organic assistant responses. However, the platform is closed: APIs and prices are not mentioned in sources, it is impossible to reproduce ad insertion within one's own dialogue product, and dependence on someone else's inventory remains the main risk for startups in this layer. Until measurement tools mature, it is reasonable to launch budgets as pilots through intermediaries or proprietary analytics, without positioning ChatGPT Ads as a primary channel.

Why this matters for users

In the 26 new countries, including Ukraine, Argentina, Chile, Colombia, and Nigeria, Free and Go plan users will begin to see ads in the chat, labeled as sponsored; according to OpenAI, they do not influence model responses. Plus, Pro, and Enterprise subscribers will continue to use ChatGPT without ads. It is worth paying attention to formats and labeling: in the US, a larger visual format is being tested for users who generate images, and the sponsorship designation is evolving from an explicit "sponsored" badge to a less noticeable side mark. The less noticeable the labeling, the more carefully users will have to distinguish where the assistant's response ends and the paid message begins.

What is currently unknown / limitations

OpenAI's claim that advertising does not influence model responses is not yet supported by methodology: there are no independent measurements of the impact of advertising context on the quality, neutrality, or speed of responses, and the principles of separating ad ranking and generation have not been disclosed. Data on advertisers—over 100,000 connected and approximately 60,000 active—is an estimate from a The Information publication citing an analyst, not an official disclosure by OpenAI. Inventory pricing, the availability of APIs, and the distribution of impressions across 89 markets are not disclosed in the sources. The preparation of ad matching for new territories is also not described: the localization of ranking and measurements for Latin America, Africa, and South Asia remains out of sight.

Sources

Author

Look at AI, editorial team