On 24 August 2026, OpenAI expanded its ChatGPT advertising test to 31 European markets, adding product carousels to the sponsored-results format it began testing in the US on 16 January 2026. Reported revenue was up more than 25% since the start of August, approaching a $1B annualised run rate, with a shift toward cost-per-click pricing, as covered by outlets including Search Engine Land and Search Engine Journal. The short answer for organic marketers: the answer surface now has paid inventory on it, roughly 20% of ChatGPT queries carry commercial intent, and the value of being the organic recommendation just went up, because it is the position that cannot be bought.
Key Takeaways
- OpenAI began testing ads in ChatGPT on 16 January 2026, US only, logged-in users aged 18+, on the Free and Go tiers, with sponsored results placed below the answer and a stated commitment to "answer independence."
- On 24 August 2026 the test expanded to 31 European markets and added product carousels.
- Reported revenue was up 25%+ since the start of August, approaching a $1B annualised run rate, with a shift toward cost-per-click.
- Roughly 20% of ChatGPT queries are estimated to carry commercial intent, the addressable slice for advertising.
- AppsFlyer added in-app measurement support; AdRoll began piloting access for smaller advertisers.
- Some field guides circulating describe specific bidding mechanics and a tracking pixel. I could not confirm those from primary sources, so treat them as unconfirmed.
- Strategic read: paid arriving below the answer makes the organic recommendation inside the answer more valuable, not less.
- As of late August 2026 this is still a test, and formats, geography, and pricing are all subject to change.
Sponsored results sit below the answer, with product carousels added in the European expansion. The answer itself is stated to remain independent.
What actually shipped
The January 2026 starting point
OpenAI's advertising test opened on 16 January 2026 in the United States. The scope was deliberately narrow: logged-in users aged 18 and over, on the Free and Go tiers. Paid subscribers on higher tiers were not part of the test. The format placed sponsored results below the generated answer rather than inside it.
"Answer independence"
OpenAI stated that ads do not influence the content of the answer: the model composes its response, and sponsored results appear separately beneath it. This is the single most important design claim in the whole rollout, and it is worth stating plainly that it is a claim about a system users cannot inspect. Taking it at face value is reasonable; treating it as verified is not.
The August 2026 European expansion
On 24 August 2026 the test expanded to 31 European markets. The notable format addition is product carousels: a visual, shoppable row rather than a text link. Carousels signal an intent to serve retail and ecommerce inventory, not just lead-gen.
The commercial signals
Reported revenue was up more than 25% since the start of August, approaching a $1B annualised run rate. The pricing model has been shifting toward cost-per-click, which is the format advertisers already understand and can budget against: a shift MarketingProfs and other trade outlets have flagged as the point at which performance budgets start moving. Roughly 20% of ChatGPT queries are estimated to carry commercial intent, that is the ceiling on inventory, and it is a large ceiling.
Ecosystem support
The supporting infrastructure is arriving faster than the format itself is stabilising, which is usually the sign that a test is intended to become a product.
AppsFlyer added in-app measurement for the format, which matters because attribution is usually what gates budget. AdRoll began piloting access for smaller advertisers, which suggests the inventory is being opened beyond enterprise buyers faster than many expected.
What I could not confirm
Several field guides circulating in the AEO community describe specific bidding mechanics, auction types, minimum bids, quality-score analogues, and a dedicated tracking pixel. I could not confirm any of that from primary sources.
I am flagging it rather than repeating it, because the fastest way to lose credibility in a developing story is to pass along plausible-sounding operational detail that turns out to be someone's inference. If you are planning spend, verify mechanics directly with the platform or your agency rep rather than from secondary write-ups, mine included.
Why this matters for organic AEO
The answer surface is now monetised
For two years the working assumption in AEO was that AI answers were a purely organic surface: you earned your way in or you didn't. That assumption is now obsolete. There is inventory adjacent to the answer, priced per click, with measurement partners and a growing advertiser base.
But the answer itself is still unbought
Sponsored results sit below the answer. If answer independence holds, the brand named inside the response is there on merit: retrieval, training, and whatever the model has learned about the category. That is the position that money cannot currently buy, and scarcity is what gives it value.
Paid raises the floor, organic keeps the ceiling
Anyone can now appear beneath a commercial-intent answer by paying. That commoditises presence. What it does not commoditise is being the brand the model reaches for in its own voice. Expect the gap in perceived credibility between those two positions to widen, not narrow, as users learn to recognise sponsored placement.
The economics: the row below is purchasable, the sentence above is not. As of late August 2026, that is the whole strategic story.
What this changes in practice
Commercial-intent queries get more crowded
If roughly a fifth of queries carry commercial intent, that is where the ads will concentrate: "best," "vs," "buy," "pricing," "alternatives." Those are exactly the queries AEO practitioners have been targeting. Organic visibility on them is now competing for attention with paid placement directly beneath.
Informational queries get relatively more valuable
Queries without commercial intent will carry less or no advertising. Being the recommended source on the informational side of the funnel becomes a cleaner, less contested position, and it is upstream of the commercial query anyway.
Budget conversations will shift
Once measurement exists (AppsFlyer) and small advertisers can buy (AdRoll pilots), finance teams will start asking why you are investing in organic AI visibility when you could buy placement. The answer is the same as it has always been with search: paid stops the day you stop paying, organic compounds, and the two positions carry different credibility. But you should have that answer ready with numbers.
Attribution will get messier before it gets better
A user who sees your brand named in an organic answer, then clicks a competitor's sponsored carousel, then returns to you via branded search is a normal path that most attribution models will assign badly. Do not let a paid dashboard take credit for demand your organic presence created.
How to respond as an organic practitioner
1. Separate your commercial and informational prompt sets
Track them apart. Their dynamics will diverge now that only one carries ads.
2. Measure recommendation, not just citation
The organic prize is being named in the answer. Score your prompt tests on whether the model names your brand, separately from whether it links you. Those metrics already move independently.
3. Strengthen the inputs that build category knowledge
Documentation, comparison pages, third-party coverage, review presence, and community discussion are what make a model confident enough to name you. None of that is purchasable in the ad auction.
4. Prepare the internal argument now
Write down what organic AI visibility is worth to your business before someone asks you to defend it against a CPC number. The comparison will be made whether you prepare for it or not.
5. Watch the format, not just the geography
The addition of product carousels is more strategically significant than the addition of 31 markets. Format changes tell you which verticals are being courted next.
Where this sits against Google's playbook
The same pattern, compressed
Search took roughly two decades to move from clean organic results to a page where paid occupies the most valuable positions. Conversational answers appear to be running the same arc in about two years. Knowing the shape of that arc is genuinely useful: it tells you that the organic position gets squeezed gradually, not suddenly, and that the squeeze starts at commercial-intent queries.
The important difference
In classic search, paid and organic compete for the same visual real estate as a list of links. In a conversational answer, the organic outcome is a sentence in the model's voice and the paid outcome is a labelled unit below it. Those are different objects, and users read them differently. That difference is the organic practitioner's structural advantage, for as long as the placement holds.
What to tell your leadership team
Lead with the format, not the geography
"Ads expanded to Europe" is a headline. "The answer surface now carries purchasable inventory on roughly a fifth of queries, and product carousels signal retail intent" is a briefing. The second version leads to decisions.
Be explicit about what is unconfirmed
Say clearly which parts are reported (revenue trajectory, CPC shift, market count) and which parts are unverified (specific bidding mechanics, pixel behaviour). Leadership tolerates uncertainty far better than it tolerates being corrected later.
Ask for a test budget, not a reallocation
The right ask is a small experimental paid budget alongside continued organic investment, with separate measurement. Framing it as a trade-off invites a false choice.
The honest uncertainties
This is a test, not a finished product. Several things could change quickly and would change the analysis:
- Placement. Sponsored results are below the answer today. If they move above it, or into it, the organic value proposition changes substantially.
- Answer independence. It is a stated commitment. It is not externally auditable.
- Geography and tiers. Currently Free and Go tiers, logged-in adults, US plus 31 European markets. All of that is expandable.
- Pricing. The shift toward CPC is reported, not fixed.
- Mechanics. As noted, specific bidding details and pixel behaviour circulating in guides are unconfirmed.
As of late August 2026, anyone giving you a confident three-year forecast on ChatGPT advertising is guessing.
Plan for both, but do not fund one by dismantling the other. Organic recommendation is the position that compounds.
Frequently Asked Questions
When did ChatGPT start showing ads?
OpenAI began testing ads on 16 January 2026 in the US, for logged-in users aged 18+ on the Free and Go tiers, with sponsored results shown below the answer.
What changed on 24 August 2026?
The test expanded to 31 European markets and added product carousels: a visual, shoppable format alongside the existing sponsored results.
Do ads influence what ChatGPT says in its answer?
OpenAI states that ads do not affect the answer: the format is described as maintaining "answer independence," with sponsored results placed separately below. That is the platform's stated position; it is not independently auditable.
How big is the ChatGPT ads business?
Reported revenue was up more than 25% since the start of August 2026, approaching a $1B annualised run rate. These are reported figures for a product still in testing.
What share of ChatGPT queries are commercial?
Roughly 20% are estimated to carry commercial intent. That is the addressable inventory for advertising, and it overlaps heavily with the queries AEO targets.
Can small advertisers buy ChatGPT ads?
AdRoll began piloting access for smaller advertisers, which suggests the inventory is opening beyond large buyers. Availability as of late August 2026 is still limited and geography-dependent.
How is performance measured?
AppsFlyer added in-app measurement support for the format. Some guides reference a dedicated tracking pixel and specific bidding mechanics, but I could not confirm those from primary sources, verify with the platform before planning around them.
Does this make organic AEO less valuable?
The opposite, in my read. Paid placement below the answer is purchasable by anyone; being named inside the answer is not. Scarcity raises the value of the organic position.
Should I shift budget from organic AEO to ChatGPT ads?
Test paid if commercial intent matters to you, but do not fund it by dismantling organic work. Paid stops when spend stops; organic presence compounds and carries different credibility with users.
Is this rollout final?
No. It remains a test as of late August 2026, and placement, pricing, tiers, geography, and format are all subject to change.
Work with me
If you are trying to work out what an ad-supported answer surface means for your organic strategy specifically, rather than in general, you can see my work and reach me through the contact form at younusfardeen.com. I bring four-plus years of marketing experience helping edtech and startup brands grow organically, and I would rather map the uncertainty with you than pretend it isn't there.