OpenAI announced on 9 February 2026 that it had begun testing ads in ChatGPT for users in the United States on the free tier and the $8-per-month Go plan. The ads appear as sponsored placements embedded directly within AI-generated responses, matched to the subject of the conversation, past chats and previous ad interactions. Early participants including Target, Albertsons and Williams-Sonoma reported notable results, with Target citing 40% month-over-month traffic growth from ChatGPT to its site, which makes this the first major AI assistant to insert paid placements into its answers at scale. For brands that have spent the past year treating AI visibility as an earned-media problem, the February launch reframed it as a question with a media budget attached, and the May expansion into self-serve tooling and international markets confirmed that the experiment is now a product.
What ChatGPT ads look like and how they are targeted
The ads sit inline within ChatGPT's responses rather than in a separate banner or sidebar. A user researching recipes might see an ad for a grocery delivery service or meal kit embedded in the answer, and a user comparing laptops might see a retailer's sponsored placement next to the recommendation. OpenAI stated that ads are matched based on "the subject of their conversations, past chats, and previous ad interactions", with no access to user data given to advertisers, only aggregated performance metrics. The company also said that "ads do not influence the answers ChatGPT gives you", and that users can dismiss ads, view why they were shown, clear their ad interaction history, and manage personalisation settings at any time.
The placements are restricted by topic and audience. Ads are excluded from sensitive or regulated topics including health, politics and mental health, and no ads are shown to users under 18. Paid subscribers on the Plus, Pro, Business, Enterprise and Education tiers do not see ads at all, which means the ad surface is limited to the free and Go tiers. OpenAI did not disclose what share of its user base sits on those tiers, but by May 2026 the company had scaled the ads to an average of 17% of daily users, which suggests a significant reach even within the restricted surface.
Early advertiser results and the Target case
The first wave of advertisers included major retailers, and the results they reported were strong enough to make the ad format credible. Target, ranked fifth in the Top 2000 retail database, stated in a 9 February corporate blog post that it had seen 40% month-over-month traffic growth from ChatGPT to Target.com. Albertsons (18th in the same database) ran Valentine's Day seasonal campaigns, and Williams-Sonoma (23rd) tested the platform alongside its West Elm and Pottery Barn brands. All three were named in coverage by Digital Commerce 360 on 13 February, citing OpenAI's blog and direct retailer announcements as primary sources.
The 40% figure from Target is notable not only for its size but for its timing. The ads launched in early February, which means the growth claim covers a window when the format was new and the ad load was light. Whether that growth rate holds as more advertisers join the platform and the ad density increases is an open question, and one that early movers will have answered by now but that OpenAI has not published aggregate data on. For brands evaluating whether to enter the platform, the Target case is the strongest public signal available, and it points towards ChatGPT ads driving measurable traffic rather than serving as a brand exercise alone.
The self-serve ads manager and international expansion in May
OpenAI expanded the ad platform in May 2026, launching a self-serve ads manager and opening the format to new markets including the UK, Mexico, Japan, Brazil and South Korea. The self-serve platform includes CPC (cost per click) bidding alongside the original CPM model, and product feed campaigns with support for up to 1 million SKUs, which shifts the offering from a managed pilot to a scalable ad product. Coverage by Digiday on 7 and 12 May confirmed the international rollout, quoting Dave Dugan, OpenAI's head of global ad solutions, and citing anonymous ad executives familiar with the product.
Want to get recommended by AI?
Check your AI search visibility, then let the Honeyb agent write, fix, and earn what gets you recommended. Free to start.
The expansion timeline is worth noting. OpenAI tested the ad format quietly through late 2025, facing backlash for what users saw as app suggestions resembling ads, then launched formally in February 2026. Within three months it had added self-serve tooling, international markets and product catalogue integration, which is a faster build-out than many observers expected given how cautiously the company had approached monetisation previously. Anthropic, OpenAI's competitor, ran a Super Bowl ad in early 2026 mocking AI advertising integration, which gives a sense of how divisive the move was within the AI industry itself, but the speed of the rollout suggests OpenAI saw enough traction in the February data to accelerate rather than pause.
What this means for brands and AI visibility strategy
The February launch split the AI visibility question into two tracks. Before ChatGPT ads, the path to appearing in an AI answer was entirely earned: write well, get cited by reputable sources, appear in the training data or the live retrieval set, and hope the engine names you. After February, that path still exists, but there is now a paid track alongside it, where a brand can buy a placement inside the answer rather than work to earn one organically. The trade-off is obvious: the paid track is faster and more predictable, but it only reaches the free and Go tiers, and it sits in a format users may learn to skip once the novelty wears off.
For performance marketers the calculation is simpler than for brand teams. If ChatGPT ads deliver a 40% traffic lift at a CPM or CPC that beats other channels, the ad format is worth testing regardless of how it feels, and early adopters have a window before the auction matures and costs rise. For brand teams the question is stickier, because an ad inside an AI answer is not the same as being the answer, and long-term brand equity is built on the latter rather than the former. A shopper who sees your brand named organically by ChatGPT as the best option for their needs is receiving a third-party endorsement; a shopper who sees your brand in a sponsored placement is receiving an ad, and they know the difference.
The cleanest way to think about it is that ChatGPT ads are a short-term visibility tactic, not a substitute for the earned work of getting cited by the sources AI engines trust. The ads can drive traffic and conversions now, which is valuable, but they do not change the underlying task of making sure that when ChatGPT answers a buying question in your category without an ad present, your brand is the one it names. For the longer strategy on that front, the work remains the same: build citeable content, earn mentions in the sources engines draw on, and track how often you appear across engines rather than in one paid slot. And for brands wondering where they stand today before deciding whether to buy their way in, the fastest path is a free check across the main engines to see whether the organic answer names you at all.
Sources: TechCrunch (9 February), Digital Commerce 360 (13 February), Target corporate blog (9 February), Digiday (7 May, 12 May).
Sources cited in this report
