The Maze: OpenAI is opening ChatGPT advertising to businesses in 26 more countries, bringing its stated coverage to 89. The October 21 expansion includes 14 Latin American markets. More businesses can now buy space beside conversations where people weigh purchases. Whether that space reliably produces extra sales is a separate question.
Source: OpenAI. OpenAI’s illustrative interface shows a labeled ad below a separate ChatGPT answer; it does not verify delivery in any new market.
The Details:
A bigger entrance, with limits. The new markets include Argentina, Chile, Colombia and Peru, alongside countries elsewhere such as Nigeria, South Africa, Bangladesh and Pakistan. Businesses can buy through technology partners or Ads Manager, OpenAI’s self-service campaign tool. Ads appear on Free and Go, its free and lower-cost plans; Plus, Pro and Enterprise remain ad-free. OpenAI says more than 100,000 advertisers have run campaigns since launch. That counts businesses that have tried the platform, rather than those actively spending today. The announcement does not disclose retention or typical advertiser returns.
Conversation creates the opportunity; an auction sells it. OpenAI describes a separate advertising service that interprets a conversation as the person’s needs become clearer. It finds possible ads, checks targeting and available budgets, assesses relevance, then runs an auction for selection and pricing. Sometimes no ad qualifies. This means a business competes for a relevant paid placement; it does not buy control of the assistant’s answer. OpenAI says advertisements remain labeled and separate, and advertisers cannot read private conversations. These are the operator’s stated safeguards, rather than an independent audit of every placement.
The business is growing; the sales evidence is younger. OpenAI disclosed a $1 billion annualized advertising revenue pace in August. That projects a year at the prevailing pace; it does not mean $1 billion had already been earned. In October it expanded measurement partnerships and described early work on geographic experiments to establish sales caused by ads. Recording a purchase after an ad and proving that the ad generated an extra purchase answer different questions. The country expansion also should not be confused with universal access to every format: OpenAI’s image-generation ad announcement described an initial US test, while business-sponsored conversations were separately being tested with selected US advertisers.
MarketMaze Take: OpenAI has widened the shop window faster than it has publicly established what a typical merchant earns from it. That is an opportunity for controlled experiments, rather than evidence for moving a fixed share of every advertising budget. Measurement investment is progress. Country counts cannot finish that job.
The competition is already conversational. Google began testing ads in its conversational AI Mode in 2025 through eligible existing campaigns, and described further format tests in May 2026. For brands across markets, the useful comparison is which buying route produces additional customers at an acceptable cost. ChatGPT adds a separate account or partner route; Google’s original tests drew on existing campaigns. Neither route’s availability alone establishes better returns.
Keep trial and loyalty apart. The 100,000 figure shows cumulative interest, while the annualized revenue milestone shows OpenAI can monetize some of that demand. Neither reveals how many businesses repeat their spending. Strong results from individual advertisers can coexist with uneven results elsewhere. The missing retention and representative sales evidence should constrain confidence, without being mistaken for proof that the channel has failed.
