OpenAI Puts Ads Inside ChatGPT. The Trust Contract Just Changed.

OpenAI began testing advertisements inside ChatGPT in February 2026, starting with Free and Go tier users in the United States. Early brand partners include Expedia, Best Buy, Enterprise Mobility, and Qualcomm, with agency groups WPP, Omnicom, and Dentsu brokering deals. CPM pricing starts at 60 dollars, three times Meta's rate, with a 200,000-dollar minimum ad spend. Paid tiers including Plus, Pro, Business, Enterprise, and Education remain ad-free.

How the Ads Work

Sponsored content appears outside ChatGPT's responses, clearly labeled and visually separated. Targeting is context-based: ads match the topic of the current conversation. Someone researching recipes might see grocery delivery ads. OpenAI also offers optional personalized advertising that draws on past chats, ad interactions, and saved memories. Sensitive categories including health, mental health, and politics are excluded. Users can opt out of personalized ads and delete all ad history data.

The design choice to place ads outside the response rather than inside it is a deliberate trust signal. OpenAI is trying to maintain the perception that ChatGPT's answers are not influenced by advertisers, even as advertisers pay to be adjacent to those answers. Whether users sustain that distinction over time is the open question.

Early Scale Signals

Analysis firm Adathena tested over 500 ChatGPT prompts and found ads appearing in just 0.8 percent of responses, confirming OpenAI is prioritizing restraint over saturation. WPP estimates 500 to 800 million dollars in first-year ad revenue, below OpenAI's internal multi-billion-dollar target but strong for a brand-new format.

The low ad density is strategic. If OpenAI saturated early, it would give enterprise and business-tier users a reason to question whether the platform's incentives had shifted. By keeping ads sparse on free tiers and absent on paid tiers, OpenAI is attempting to run a dual-business-model company: subscription for high-value users, advertising for scale users.

The Trust Calculus

The fundamental tension is structural. An AI assistant's value comes from the user's belief that its recommendations are objective. Advertising, by definition, introduces a sponsor's interests into the information environment. OpenAI's design choices mitigate this, but they do not eliminate it.

For enterprise buyers evaluating ChatGPT as a productivity tool, the relevant question is not whether free-tier ads affect their paid deployment. It is whether the existence of an advertising business model changes how OpenAI prioritizes product development. When ad revenue becomes material, the features that drive ad impressions, such as consumer engagement and session length, compete for engineering resources with the features enterprise buyers care about, such as accuracy, reliability, and data governance.

What This Means for Marketing Budgets

At 60 dollars CPM and 200,000-dollar minimums, ChatGPT advertising is enterprise-grade media buying, not experimental budget. The brands moving first are retail, software, and travel, categories where ChatGPT is already a primary research tool for consumers.

The unit economics are worth scrutinizing. A 60-dollar CPM means 60 dollars per thousand impressions. If ChatGPT's ad placement drives even moderate conversion rates, the cost per acquisition could undercut display and social channels where CPMs are lower but intent is weaker. The comparison that matters is not CPM-to-CPM. It is cost-per-qualified-action in a channel where the user has explicitly stated what they want.

The Competitive Framing

Google's approach monetizes through AI Mode product placements where the ad is embedded in the answer. OpenAI's approach keeps ads visually separate from the answer. Both are competing for the same marketer budgets. Google has the advantage of existing advertiser relationships and the Shopping Graph. OpenAI has the advantage of 900 million weekly active users who trust the platform enough to ask it personal questions.

The brands that test now will learn whether conversational AI is a high-intent channel worth scaling. The brands that wait will be buying at higher CPMs in 2027 after the early data proves the format works.

What Could Go Wrong

If user satisfaction surveys show that even clearly labeled ads reduce trust in ChatGPT's recommendations, OpenAI faces a choice between revenue growth and product integrity. Klarna's experience with AI deployment, where cost savings initially looked compelling but customer satisfaction declined, is an instructive parallel. The other risk is regulatory. If the FTC determines that contextual ads in AI assistants require different disclosure standards than web ads, compliance costs could erode the margin advantage that makes the format attractive to advertisers.