OpenAI embraces the inevitable: ads come to ChatGPT

Fata viam invenient. -Virgil

OpenAI announced today that it will soon (“in the coming weeks”) begin testing advertising in its Free and Go tiers in the United States. From its announcement (emphasis mine):

In the coming weeks, we’re also planning to start testing ads in the U.S. for the free and Go tiers,
so more people can benefit from our tools with fewer usage limits. Plus, Pro and Enterprise
subscriptions will not include ads … People trust ChatGPT for many important and personal tasks, so as we introduce ads, it’s crucial we preserve what makes ChatGPT valuable in the first place. That means you need to trust that ChatGPT’s responses are driven by what’s objectively useful, never by advertising. You need to know that your data and conversations are protected and never sold to advertisers. And we need to keep a high bar and give you control over your experience so you see truly relevant, high-quality ads — and can turn off personalization if you want.

Notably, OpenAI also announced today that it is expanding its Go tier — its lowest-priced paid tier at $8/month — to be available globally.

I’ve long maintained that ads in ChatGPT were inevitable. From Obviously, OpenAI will monetize with ads, which was published in May 2025:

Obviously, OpenAI will monetize with ads. Hiring Simo represents such an on-the-nose acknowledgment of that fact that I almost didn’t write this piece. Except that OpenAI’s admission that advertising is its path forward on monetization serves to dispel a common misconception (really, a fallacious, superstitious tech dogma) that advertising is but one of many monetization strategies that are all equally capable of achieving optimal revenue for scaled consumer technology products. This isn’t true. If maximizing revenue is an organization’s objective function, and its product can potentially reach a scale of billions of users, then advertising stands alone as its optimal monetization strategy.

I make the case in Affiliate links, personalized ads, and chatbot revenue optimization that OpenAI necessarily must introduce ads to ChatGPT because all other monetization models are simply sub-optimal for a product like ChatGPT with a humanity-spanning TAM (currently sitting at 900MM WAU). One of the principal arguments I’ve seen made against the introduction of ads to ChatGPT, and chatbots more generally, is that contextual ads could be seen as undermining the trust that users place in the responses delivered by chatbots, since the chatbot operator could be seen as serving two potentially divergent interests. In other words: that users wouldn’t be able to trust chatbot responses as anything more than a delivery mechanism for a sales pitch.

I’ve made the same argument myself, classifying those potentially competing interests as an AI-search incentive problem. But as I argue in Affiliate links, personalized ads, and chatbot revenue optimization, as with social media, there is no reason that ads must be contextually-targeted; if they are targeted against behavioral profiles, they might be wholly unrelated to the core context of a chatbot discussion. From that piece:

But consumer-facing products that have humanity-spanning appeal — which, at 800MM WAU, ChatGPT does — are not limited to contextual signals as inputs for advertising targeting. And as I alluded to above, neither is Search, so even when the chatbot use case is seen as qualitatively similar to that of Search, the imposition that only contextual data may be used for advertising targeting is an unnecessary restriction.

Ben Thompson and I discussed this point at length in our recent conversation on his podcast:

ES: Right, and they don’t have to be. You build a CAPI [Conversions API], you build a pixel, you collect data from advertisers — the advertisers want you to have this data, they want you to have the conversion data because then you can do targeting on that basis, they want you to have it. And so, the way you sidestep that editorial problem, the conflict-of-interest problem, you sidestep the issue of the user questioning whether the output is serving the purpose of framing the ad or vice versa, is you don’t show ads related to the context at all … Contextual ads don’t work as well as personalized ads, what you do is you take data on the back-end from the advertisers that have sold stuff to you, you use that to determine what the right ad to show to that person is, and you put it alongside the content. You put it between content blocks, but if it’s unrelated to the actual content, then no one ever questions whether you’re making the decision on what goes in the content on the basis of who’s advertising against it. So, you side-step that conflict issue altogether.

To that end, OpenAI outlined a number of principles for its advertising product in the blog post that reveals it. One of those principles is Answer independence, which the blog post characterizes with:

Ads do not influence the answers ChatGPT gives you. Answers are optimized based on what’s most helpful to you. Ads are always separate and clearly labeled.

This approach doesn’t go as far as making ads utterly unrelated to the context of the question, but it does decouple the chatbot’s content from ad selection beyond context. The blog post goes on to state that, “The best ads are useful, entertaining, and help people discover new products and services.” This suggests targeting that will be supported by behavioral data, and I’d expect ChatGPT to introduce a CAPI and / or pixel for that purpose.

I believe that ChatGPT’s advertising platform will be wildly successful. The gap between OpenAI’s and Meta’s unit economics is attributable to the latter’s performant, direct-response advertising platform. And while I also believe that OpenAI is late with its advertising offering, given that Alphabet has already found a means of monetizing its Gemini model without inserting ads into the Gemini chatbot, I don’t think that puts OpenAI at any sort of material disadvantage or jeopardizes the project. Ultimately, ads in ChatGPT will enable OpenAI to monetize the product at the scale necessary to support the potential of its adoption. If anything, ads in ChatGPT are a gift to its user base, which, as the blog post points out, will now benefit from “fewer usage limits.”