OpenAI Opens ChatGPT Ads to Self-Serve with CPC Bidding and Conversion Optimization

When Google launched AdWords in 2000, it made a bet that felt counterintuitive at the time: that placing ads inside a search engine would feel helpful rather than intrusive, because users who are searching are actively looking for something. OpenAI is making the same bet about conversational AI — and on June 5, it took a significant step toward proving it can build a commercial engine as powerful as the one it’s threatening to disrupt.

The company rolled out conversion-optimized campaigns across its Ads Manager platform, extended self-serve access to U.S. advertisers without a minimum spend threshold, and added cost-per-click bidding alongside the existing cost-per-impression model. The changes together represent a maturation of OpenAI’s advertising business from brand experiment to full-stack performance marketing platform.

What Actually Changed on June 5

Three concrete updates went live:

Conversion-optimized campaigns. This is the most significant change for performance marketers. Until now, ChatGPT ad campaigns were optimized for awareness (impressions) and traffic (clicks). Conversion optimization shifts the objective to downstream actions — purchases, form fills, app installs, lead submissions — the metrics that performance marketing budgets actually care about. Without this capability, ChatGPT was useful for brand awareness spending but largely unusable for direct response advertisers, who control the majority of digital ad dollars.

Cost-per-click bidding. Advertisers can now choose between paying per impression (CPM) or per click (CPC). CPC aligns ad spend more directly with measurable engagement and lowers the barrier for smaller advertisers who need to manage budgets against performance rather than reach.

Removal of the minimum spend threshold. During the pilot phase, advertisers were required to commit at least $50,000 to test campaigns. That floor is gone. Any business can now register in the Ads Manager, add payment details, set a budget, upload creative, and launch — the same frictionless self-serve model that made Google AdWords and Facebook Ads Manager accessible to small businesses in the 2000s and 2010s.

The Infrastructure Behind It

Alongside the self-serve expansion, OpenAI has deployed a measurement framework that attempts to resolve one of the longest-standing debates in digital advertising: how to attribute outcomes in a privacy-sensitive environment.

The platform uses a dual-track approach: a Conversions API for server-side event tracking (matching purchase signals from advertiser backends to ad exposure without transmitting user-level conversation data) and a pixel-based measurement option for web-based conversion tracking. Critically, OpenAI has committed that ad placement and performance data will not be derived from the content of individual user conversations — a line it has publicly maintained since the advertising program began.

“We’re creating a new ads model — one that supports businesses and broader access to AI while staying grounded in clear principles around answer independence, privacy, and user control,” said David Dugan, Head of Global Solutions at OpenAI.

The phrasing “answer independence” is doing real work here. It addresses the concern that surfaced the moment OpenAI announced an ad business: that ads would distort the AI’s responses. OpenAI’s stated commitment is that ads occupy defined placements, do not influence the model’s substantive answers, and are disclosed as advertising. The actual enforcement of that boundary — and whether users perceive it as credible — remains the most important variable in whether ChatGPT advertising achieves durable user trust.

The Revenue Stakes

OpenAI is targeting $2.5 billion in advertising revenue in 2026 and has reportedly set a longer-horizon goal of $100 billion annually by 2030. Those figures, if accurate, would make ChatGPT one of the largest advertising platforms in the world within four years — comparable in scale to Meta’s 2019 ad revenue.

The addressable opportunity is substantial. ChatGPT reportedly handles over 500 million weekly active users as of mid-2026, a user base that skews toward higher-income, technology-adjacent demographics that advertisers pay premiums to reach. More importantly, ChatGPT users engage during active decision-making: researching purchases, comparing options, solving problems. That behavioral context is structurally more valuable for advertising than passive content consumption on social media.

The agency ecosystem is already onboard. Dentsu, Omnicom, Publicis, and WPP are signed as agency partners, giving OpenAI reach into the major holding companies that control a substantial fraction of global enterprise ad spending. Technology partners including Adobe, Criteo, Kargo, Pacvue, and StackAdapt expand the programmatic infrastructure that large-scale advertisers require.

What This Means for Google and Meta

The launch of conversion optimization on ChatGPT should be read as a direct competitive move against Google’s Performance Max campaigns and Meta’s Advantage+ shopping campaigns — the products that dominate performance marketing for e-commerce and lead generation advertisers.

Google’s advertising moat has always rested on two foundations: the volume of commercial intent signals from search queries, and the feedback loop of conversion data that makes its optimization algorithms increasingly accurate over time. ChatGPT’s queries carry comparable or superior commercial intent signals — users describing their purchasing problem in natural language often provide more context than a three-word search query. The key question is whether OpenAI can accumulate enough conversion data fast enough to make its optimization algorithms competitive.

Meta’s moat is different: social graph data and remarketing precision across Facebook and Instagram. ChatGPT doesn’t have that. But it doesn’t need to replace Meta — it only needs to capture a portion of the performance marketing budget that advertisers increasingly diversify across platforms.

The Small Business Inflection Point

The removal of the $50,000 minimum is potentially the most consequential change for the medium-term trajectory of ChatGPT advertising. Large brand advertisers and agency-managed campaigns could participate with high minimums. But the self-serve, low-barrier model that scaled Google and Facebook was fundamentally driven by the long tail: millions of small and medium-sized businesses spending $500 to $5,000 per month.

Those advertisers now have direct access to ChatGPT’s audience. The interface is familiar enough — budget setting, creative upload, bidding strategy — that any marketer comfortable with Google Ads or Meta Business Manager can onboard without training. The creative format is the variable: ads in conversational AI must fit into a dialogue context, which requires different copy approaches than search ads or social banners.

The Unknowns

Several important questions remain unanswered. OpenAI has not publicly disclosed the click-through rates, cost-per-click benchmarks, or conversion rates that early pilot advertisers have achieved. Without those figures, it’s impossible to assess whether ChatGPT ads generate comparable ROAS (return on ad spend) to established platforms.

The geographic rollout is currently U.S.-only, with expansion to the U.K., Brazil, and Japan reportedly in planning. International markets will require both regulatory compliance with local advertising standards and localized measurement infrastructure.

And the fundamental model tension — between a product that users trust for objective answers and a platform that earns revenue from advertisers who want to influence those users — has not been definitively resolved. It is the central challenge OpenAI will have to navigate for as long as it runs both businesses under the same brand.

Whether it succeeds will define one of the most consequential business stories of the next five years.