OpenAI has been building toward this moment since February 2026, when ChatGPT ads launched in a closed pilot with a handful of large brand partners and a $250,000 minimum commitment. Three months later, the minimum is gone, the platform is open to any US advertiser, and the buying model has expanded from CPM-only to include cost-per-click.
This is a significant structural moment, not because ChatGPT advertising is proven at scale, but because the infrastructure required to make it a serious performance channel has now been built. Self-serve access, click-based bidding, pixel measurement, and a Conversions API are not incremental features. They are the table stakes that allow performance marketers to evaluate a channel seriously rather than as an experiment.
Here is what was announced, what the numbers look like, and what an honest assessment of the opportunity looks like for e-commerce brands right now.
What Actually Changed On 5 May
Self-serve platform
Beta Ads Manager: open to all US advertisers
Accessible at ads.openai.com after account verification. Advertisers can set budgets and bids, upload creatives, control pacing, launch campaigns, and view performance directly. No agency relationship required. No minimum spend. Previously, the entry point was a $50,000 minimum that dropped from an opening rate of $250,000; now it is zero. Available to SMBs, startups, and global brands alike.
New buying model
CPC bidding alongside existing CPM
Previously, ChatGPT ads were sold exclusively on a cost-per-thousand-impressions basis, with CPM rates falling from an opening rate of $60 to around $25 in some transactions. CPC bidding is now available, with recommended starting bids between $3 and $5 per click. Both CPM and CPC remain available. For performance marketers, CPC is the model that makes serious evaluation possible. You can align spend with user actions rather than impressions in an environment where volume is still limited.
Measurement
Pixel tracking and Conversions API
Advertisers can now measure purchases, sign-ups, and leads attributed to ad interactions through pixel-based tracking and a Conversions API. OpenAI states that only aggregated data is shared, not the content of individual conversations. Cost-per-action bidding and third-party measurement have been confirmed as in development, with no timeline given for either.
Partner ecosystem
Agency holders and ad tech platforms
Advertisers can buy through OpenAI’s Ads Manager directly or through agency and technology partners. Confirmed partners include Dentsu, Omnicom, Publicis, and WPP on the agency side, and Criteo, Adobe, Kargo, Pacvue, and StackAdapt on the ad tech side. This means ChatGPT inventory will begin appearing in existing buying workflows and you may be able to buy it through tools you already use before the end of the year.
How Fast This Moved: The Timeline
The speed of OpenAI’s ad platform build is worth understanding.
The infrastructure that typically takes established platforms years to build has been assembled in under four months.
Feb 2026
Pilot launch: Closed roster of large brand partners. CPM-only buying. Minimum commitment $250,000. Reported $100m+ in revenue after six weeks.
Early 2026
Pilot expansion: Minimum drops to $50,000. CPM rates fall from $60 to ~$25. CPC bidding was introduced to pilot advertisers.
5 May 2026
Self-serve launch: Ads Manager beta opens to all US advertisers. Minimum spend dropped entirely. CPC bidding broadly available. Pixel and Conversions API live. Partner ecosystem confirmed.
TBC 2026
Next phase: Cost-per-action bidding and third-party measurement are both confirmed as in development. No partners or timeline disclosed.
- $2.5B: OpenAI's ad revenue target for 2026
- $3–5: Recommended starting CPC bid range
- $100B: OpenAI's longer-range ad revenue ambition by 2030
The Honest Assessment For eCommerce Brands
The infrastructure announcement is real and significant. The performance case for e-commerce is still being built. Those are two different things, and it’s worth holding both clearly.
What’s genuinely compelling
ChatGPT's user base is large, engaged, and skewed toward decision-oriented queries. Many ChatGPT conversations are active and decision-oriented, which is exactly the context in which advertising can add value rather than interrupt. A user actively researching a purchase decision inside ChatGPT is, in intent terms, at least as valuable as a high-commercial-intent search query.
The early CTR data is encouraging. Reports from the pilot phase describe strong engagement rates relative to established display channels — though the small scale and premium placement mean those numbers should be read carefully rather than extrapolated.
The partner ecosystem matters. Advertisers can create ads through the tool directly or through agency partners such as Dentsu, Omnicom, Publicis and WPP, as well as ad tech companies such as Adobe, Criteo, Kargo, Pacvue and StackAdapt. For e-commerce brands already using any of these platforms, the path to testing ChatGPT inventory is shorter than it might appear.
The case for testing now
Early movers on new channels consistently benefit from lower CPCs and less competition before the platform matures. The recommended $3–5 CPC starting bid is meaningfully lower than equivalent intent-based clicks on Google Search for most e-commerce categories. If you have a budget for channel experimentation, the infrastructure is now solid enough to generate real data rather than anecdote.
What needs watching
Scale remains the open question. ChatGPT has hundreds of millions of users, but the volume of ad impressions available within those sessions, and how that volume is distributed across categories and query types, is not yet publicly documented. Strong CTRs on limited inventory are not the same as a scalable channel.
Third-party measurement is not yet live. Cost-per-action bidding and third-party measurement are both in the works, but OpenAI declined to specify when either would arrive or which outside partners might be involved. Until third-party measurement is available, you are relying on OpenAI's own attribution, which creates the same measurement conflict that exists with every walled garden. The Conversions API and pixel are a start, but independent validation is the standard sophisticated advertisers will require before committing serious budget.
Category restrictions apply. OpenAI has not published a comprehensive list of restricted categories, but the platform is not open to all advertising verticals. Worth confirming your category qualifies before investing time in setup.
The questions to answer before scaling
What’s the actual impression volume available in your product category? What does attribution look like in practice? How is an ad interaction within a ChatGPT conversation connected to a downstream purchase? How does the CPC compare to your Google Search CPC for equivalent intent queries? These are answerable questions, but only through live testing, not through extrapolating from the pilot reports of other advertisers in different categories.
The Bigger Picture: What This Means For Paid Search
ChatGPT’s ad platform launch is the clearest signal yet that the answer itself is becoming the ad placement. When a user asks ChatGPT for a product recommendation, and the response includes a sponsored result, the traditional distinction between organic and paid, between search results and advertisements, becomes harder to maintain.
The search result and the ad are converging. ChatGPT's self-serve launch is the moment that the shift moves from theoretical to operational.
For e-commerce brands, the implication isn’t that ChatGPT replaces Google. The scale difference remains enormous. It’s that the intent-capture layer is no longer exclusively Google's to own. A user who might have gone to Google Search to research a purchase is increasingly going to ChatGPT to have the research done for them. How that session is monetised, and by whom, is now an active competitive question rather than a hypothetical one.
Google is watching this closely. The timing of ChatGPT's self-serve launch, landing in the same week as Google Marketing Live, is not accidental. Both companies are making the case, simultaneously, that their platform is the right place to capture intent at scale. The next twelve months will produce real data on which argument holds.
Our Honest Read
The infrastructure is now serious enough to be tested. The performance case for e-commerce at scale hasn’t been made yet, because the scale hasn’t been there to make it.
The brands that will have the most useful data twelve months from now are the ones that start generating it today, with realistic budgets and clear measurement frameworks, rather than waiting for the channel to prove itself on someone else's budget first.
ChatGPT ads have moved from experiment to platform in under four months. That's fast by any standard. The self-serve launch, the drop of the minimum spend, the CPC bidding model, and the partner ecosystem together represent a channel that is now structurally comparable to established paid media platforms, even if the scale and measurement maturity aren’t there yet.
It’s worth paying attention to. It’s worth testing with an appropriate budget. It is not yet worth reallocating significant spend away from channels with proven scale and attribution. Both of those things can be true at the same time.
Need a fresh perspective? Let’s talk.
At 360 OM, we specialise in helping businesses take their marketing efforts to the next level. Our team stays on top of industry trends, uses data-informed decisions to maximise your ROI, and provides full transparency through comprehensive reports.
.png)
