OpenAI’s ChatGPT Ads business has reached a $1 billion annualized revenue run rate less than 200 days after launch, yet it is expected to fall short of the $2.4 billion in 2026 advertising revenue the company projected to investors.
According to recent reporting, the number of advertisers roughly tripled to about 60,000 between late August and early October, driven in part by price reductions. One agency founder cited costs of roughly $5 per click for certain B2B clients, compared with $30–$50 on Google for comparable campaigns. Ads now run in more than 60 countries, and OpenAI has added visual ad formats and expanded self-serve access through Ads Manager.
The advertising team has grown to around 140 people, with a significant share previously at Meta, and has been reorganized multiple times. Leadership includes Dave Dugan, a former Meta executive. OpenAI has also worked with major holding companies and measurement partners. Early advertiser feedback is mixed: some report strong returns and new customer acquisition, while others note relevance issues, with independent analysis finding that a portion of placements have limited connection to the query.
Ads appear only for Free and Go tier users, are labeled, and do not influence ChatGPT’s answers or give advertisers access to private conversations. The platform has also tested sponsored agents that allow users to continue a conversation with a brand’s AI after seeing an ad.
For the broader advertising market, ChatGPT Ads represent a new high-intent conversational channel that can capture demand earlier in the research process. At current scale it remains a small share of total digital ad spend and faces the same measurement and targeting challenges that have slowed other emerging formats. If OpenAI continues to improve relevance, measurement and pricing efficiency while expanding formats and markets, the channel could exert incremental pressure on traditional search and social budgets. For now, most advertisers are still treating it as a test rather than a core channel.
Sources: OpenAI, The Information, New Street Research, industry reporting
