ChatGPT ads are here: what they mean for getting cited

The moment an answer engine starts selling ad space, a question follows immediately: if I can pay to appear, why work to be cited? OpenAI started testing ads in ChatGPT on February 9, 2026, opened a beta self-serve Ads Manager on May 5, and on August 31 said the ad business had reached a $1 billion annualized revenue run rate. The answer to that question is more interesting than it looks. Paid placement does not replace being named. It makes being named matter more.

Search engines went through this exact transition two decades ago. Results were organic, then ads appeared above and beside them, and a whole discipline grew up around the difference between earning a spot and buying one. AI answer engines are now at the same threshold, and ChatGPT has already taken the first big steps.

What OpenAI launched, and when

ChatGPT advertising arrived in stages rather than on a single launch day:

• February 9, 2026. OpenAI began testing ads in ChatGPT in the US, for logged-in adult users on the Free and Go tiers. The Plus, Pro, Business, Enterprise and Education tiers do not have ads.
• May 5, 2026. OpenAI launched a beta self-serve Ads Manager, where advertisers can create ads themselves, and added cost-per-click (CPC) bidding alongside the cost-per-impression buying of the first phase.
• August 11, 2026. ChatGPT Ads launched in the United Kingdom, Mexico, Brazil, Japan and South Korea.
• August 31, 2026. OpenAI said ChatGPT Ads had reached a $1 billion annualized revenue run rate in less than 200 days after launch, with tens of thousands of advertisers, and opened self-serve access across India, Europe, the Middle East and North Africa.

OpenAI says ads are always clearly labeled as sponsored and visually separated from the organic answer, and that ads do not influence the answers ChatGPT gives. For the test, it said ads would not be shown to users it knows or predicts to be under 18, and would not be eligible to appear near sensitive or regulated topics like health, mental health or politics. The rules have since loosened for some advertisers: Marketing Brew reported in June 2026 that OpenAI's ad policy no longer categorically blocks medical, legal and financial contexts, which instead need case-by-case approval.

The design choice is telling. The ad is marked as sponsored and kept visually distinct from the response itself. OpenAI is drawing a line between the answer and the advertisement, at least for now.

"An ad buys you a labeled box. A citation is the AI presenting you as part of its own answer. Those are not the same thing."

Curious whether AI engines name you when buyers ask about your category? Get a free visibility audit and see where you stand across ChatGPT, Gemini, Perplexity and Google AI Overviews.

Why this makes organic citation more valuable, not less

The intuitive fear is that ads devalue the work of getting cited. Why earn a mention if you can buy one? But look at how the two placements actually read to a user.

A sponsored card is likely to be read as advertising. People can apply the discount they apply to all advertising: this brand paid to be here. A citation inside the answer reads completely differently. It is the AI, in the course of answering, presenting your brand as part of what it genuinely found and trusts. One is rented attention with a label attached. The other is the engine vouching for you.

As answers start carrying paid slots, that distinction gets sharper, not blurrier. The earned mention can become the credible one because the paid one exists next to it for contrast.

Ads are not a rare sight, at least in some samples. OtterlyAI tracked a fixed set of commercial (shopping) prompts across 16 US industries daily from July 2 to 15, 2026, on ChatGPT only, and found an ad on 76.4% of commercial answers. In 12 of the 14 industries where Otterly named a market leader, the market leader ran no ChatGPT ads. Otterly does not disclose its prompt or answer counts, so treat 76.4% as one study's sample rather than a general frequency.

The real risk to watch

There is a legitimate worry underneath the optimism, and it is worth naming. If, over time, the line between sponsored and organic content in AI answers erodes - if paid influence seeps into what looks like the earned answer - then the value of a citation erodes with it. That is the scenario brands and regulators should watch, and it is part of why the current clear labeling matters so much.

For now, the separation is explicit. The strategic response is not to panic about ads, but to make sure your presence in the earned part of the answer is strong, so that whatever happens to the ad layer, you are established where trust lives.

Ads or citations: how to think about it

• Different jobs. Ads can buy immediate, labeled reach. Citations earn credibility a label cannot. Many brands may need the second more than the first.
• Citations are the durable asset. A paid slot lasts exactly as long as you pay. Being the source an engine reaches for can compound over time, and it does not switch off when the budget does.
• Do not let ads become an excuse. The temptation will be to buy visibility instead of earning it. A brand that is never named organically is renting its presence in AI, not owning it.
• Watch the labeling. The value of your citation depends on answers keeping earned and paid content visibly distinct. It is worth paying attention to how that holds up.

The takeaway

ChatGPT selling ads is not the end of organic AI visibility; it is the event that clarifies its worth. When answers carry paid placements, the earned citation can become the trusted one by contrast - the part of the answer the AI presents as its own. Brands should treat the arrival of ChatGPT ads not as a reason to reach for the ad budget, but as a reason to make sure they are named where it counts: inside the answer, on merit, where a sponsored label will never appear.