AI & the open web

ChatGPT Ads Get Real Plumbing — and Product Carousels Aimed at Q4

APH Marketing Desk ·5 min read Share Print
In this piece
    Figure The four points the piece sets out Why this matters for publishers →
    1. 01

      This competes for search budgets, and search budgets fund your referral economy.

    2. 02

      Commerce content faces a structural, not cyclical, threat.

    3. 03

      The Q4 timing forces the issue this year.

    4. 04

      The ambition is stated even if the numbers aren't proven.

    Marketing Desk

    Two announcements in the same week, from two directions, describe one platform growing up fast. OpenAI now has CPC bidding, pixel tracking, conversion APIs and low spend minimums — the plumbing of a real ad platform, with a barrier to entry low enough that any mid-sized advertiser can test. And days apart, OpenAI added product carousels to ChatGPT ads: multiple products in one placement, pulled straight from retailer feeds, with OpenAI’s systems deciding which products appear and whether to run a single unit or a carousel — currently one retailer per carousel, building on the bulk product-feed campaign tools shipped three months ago. There are no ad units to build at all.

    What the platform still lacks is the thing buyers ultimately pay for: attribution anyone trusts. AI Digital CSO Mary Gabrielyan was blunt: “We don’t have that answer for ChatGPT ads yet.” Until buyers can model cross-channel influence and offline outcomes, big money stays cautious.

    The timing of the carousel launch is not subtle. One agency executive said “OpenAI wants to make sure that these products are discoverable in ChatGPT conversations” ahead of the holiday season, and Adthena’s Ashley Fletcher notes Q4 “will be key” as advertisers set 2027 budgets. Behind it sits reported ambition on an uncomfortable scale: $2.5 billion in ad revenue in 2026, $100 billion-plus by 2030. Those are reported targets, not results — but they tell you what OpenAI is building toward, and how fast it intends to get there.

    01Read it as search, because that’s what it is

    The classification question matters more than any feature. ChatGPT ads do not behave like display, social or CTV — they behave like intent-driven search: a user expresses a need in a query-shaped conversation, and an ad answers it. That single fact determines whose budgets this platform competes for. Not brand video money, not social feeds — the performance budgets attached to commercial intent. Which is to say, the budgets attached to the queries that used to send publishers referral traffic.

    For publishers, that creates two distinct exposures, and they deserve separate ledgers.

    The first is budget-side and gradual: as intent moves from traditional search into AI conversations, the money follows the intent, and the referral traffic that intent used to generate — and the ad revenue riding on it — erodes with it. No publisher can stop that migration. Every publisher can measure their exposure to it.

    The second is sharper and belongs specifically to commerce publishers. The carousel pulls straight from retailer feeds. Look at what is absent from that sentence: the review site, the gift guide, the “best X for Y” page. The product-discovery step that commerce publishers monetise through affiliate links is being rebuilt as a conversation between a user, an AI and a retailer’s feed — with no publisher layer in it. Google Shopping, whatever its faults, coexisted with an ecosystem of publisher commerce content feeding and surrounding it. This run at Google Shopping is being assembled without that layer, ahead of the biggest commerce quarter of the year.

    And the attribution gap that keeps budgets cautious today is a temporary comfort. Measurement gaps get closed — pixel tracking and conversion APIs are precisely the infrastructure for closing them. Plan for the platform that closes it, not the one that exists today.

    02Why this matters for publishers

    This competes for search budgets, and search budgets fund your referral economyBecause ChatGPT ads behave like intent search, every dollar that migrates there is a dollar attached to queries that once produced publisher clicks. The competition is indirect but aimed at the foundation.
    Commerce content faces a structural, not cyclical, threatAffiliate revenue depends on publishers owning a step in product discovery. A carousel fed directly by retailer feeds deletes that step rather than competing with it. If commerce content is a meaningful revenue line, this is aimed at you specifically.
    The Q4 timing forces the issue this yearCarousels shipping ahead of holiday season, with agencies calling Q4 decisive for 2027 budgets, means the first real data on how much product discovery moves off the open web arrives in the next two quarters — not at some comfortable future date.
    The ambition is stated even if the numbers aren't proven$2.5 billion this year and $100 billion-plus by 2030 are reported targets. Treat them as claims — but as claims about intent, they are perfectly clear.
    Separately, these are two product updates; together they are a platform declaring its category.

    03What publishers should do

    04The bottom line

    Separately, these are two product updates; together they are a platform declaring its category. CPC bidding, conversion APIs and low minimums are search-platform plumbing. Feed-driven product carousels timed for Q4 are a search platform’s commercial endgame. The attribution hole is real, and Gabrielyan is right that buyers cannot yet defend the spend — but plumbing exists to be connected, and the gap between “accessible” and “measurable” is where every major ad platform in history has spent roughly one product cycle. Publishers should use that same cycle deliberately: baseline the traffic, stress-test the commerce lines, and negotiate a place in the feed-driven discovery layer while there is still leverage to negotiate with. The platform that exists today is easy to dismiss. The one that exists after the measurement gap closes will not ask permission.

    Sources & caveats

    Sources: AdExchanger, “ChatGPT Ads Are Here. Now Comes the Hard Part.” (August 7, 2026), including the Mary Gabrielyan quote and the platform-infrastructure details; Digiday, “OpenAI brings product carousels to ChatGPT ads” (August 6, 2026), including the carousel mechanics, the unnamed agency executive quote, the Ashley Fletcher quote and the Q4 timing. The $2.5 billion 2026 and $100 billion-plus 2030 ad revenue figures are reported ambitions attributed to OpenAI, not disclosed results or confirmed targets. The scenario modeling of discovery loss is the author’s operator analysis, not a projection from either source.

    The weekly

    One letter a week, from the desk that runs the auctions.

    What actually moved in yield, CTV and curation across our publishers — written by the people who saw it, not a content team. No digests, no roundups, one email.

    One email a week. Unsubscribe in one click. We never share or sell the list.

    More from this issue

    Ran alongside this piece in the Weekly of 9 August 2026 — read the whole issue →