What are ChatGPT ads?
ChatGPT ads are sponsored placements shown inside ChatGPT conversations, labelled as sponsored and kept separate from the answer itself. OpenAI ran a managed pilot with a small group of large brands from February 2026, then opened a self-serve Ads Manager beta to US advertisers on May 5, 2026.
Advertisers can now buy on cost per click or cost per thousand impressions, with no minimum spend. Our ChatGPT ads cost guide covers current pricing in detail.
What is confirmed, and what is speculation
Confirmed: ads run inside ChatGPT for US users, a self-serve Ads Manager exists, and you can bid per click or per thousand impressions without a minimum spend. Availability has also widened beyond the US pilot markets.
Still thin: targeting options, reporting depth, conversion measurement and attribution are far less mature than on Google or Meta, and they are changing quickly. Treat any precise claim about how ChatGPT ads perform as early evidence, not a benchmark.
Why a conversational ad surface is different
ChatGPT ads do not behave like search ads. The interaction is a conversation, not a query and a list of links, so a sponsored placement has to fit inside a dialogue where the user is asking for help, comparisons, or recommendations. That changes the nature of the unit: the value is in being the recommended answer, not in occupying a slot above ten links.
It also raises hard questions the industry has not resolved. How is a sponsored recommendation disclosed? How is performance measured when there may be no click in the traditional sense? How does attribution work across a conversation? These are not edge cases; they are the whole model, and they are unsettled.
How ChatGPT ads compare to Google Ads
The honest comparison is between a mature ad platform and a young one. Google Ads has years of formats, targeting controls and conversion measurement you can plan a quarter around. ChatGPT Ads Manager has the basics of buying and reporting, but far less to plan on.
The structural difference is the surface itself. A search ad sits above a list of links and is bought against a query; a conversational placement would have to sit inside a dialogue and compete to be the recommended answer. Until measurement matures, core budget stays in Google Ads and ChatGPT ads get a capped test.
What to do now (and what not to)
If your buyers use ChatGPT to research purchases, run a small, capped test: one clear offer, a landing page built for it, and success judged on your own sales or lead data rather than the platform's reporting. Keep improving how your brand is represented in the content these models draw on, because being a credible, citable answer matters whether or not you pay for placement.
What not to do is move core budget off proven channels, or trust anyone promising predictable ChatGPT ad returns this early. Early access to a new channel can be an advantage, but only if you measure it honestly.
The nearer-term opportunity: being cited by AI
There is a more actionable version of the ChatGPT question that does not depend on ads existing at all. Conversational assistants already recommend products, services, and providers by drawing on the wider web, so how your brand is represented in the content these models consume shapes whether you show up in their answers today.
This is why iClick invests in clear, factual, well-structured content, exactly the kind of glossary and guide surface this site is built on. Being the accurate, citable source an AI reaches for is a real acquisition channel in 2026, and unlike ChatGPT ads, it needs no media budget at all.
