ChatGPT ads: intent-driven, mid‑funnel inventory, not another banner slot
OpenAI’s ChatGPT advertising platform turns conversational queries with commercial intent into mid‑funnel ad inventory by placing paid messages below assistant responses for free and Go users, aiming to capture in‑progress purchase decisions without disrupting the core chat experience. At its first Cannes Lions appearance, OpenAI used a high‑profile stage to tell CMOs it is “clearly in the advertising business now,” backing that claim with a projection of USD 2.5 billion (approx. RM11.5 billion) in ad revenue and a longer‑term USD 100 billion (approx. RM460 billion) ambition. This is not a shy experiment. It is a loud attempt to turn 920 million weekly users, one‑fifth of whose queries already show commercial intent, into a new media channel that sits between search and commerce. For brands, the real story is whether that intent can be measured and scaled fast enough to justify shifting budgets away from the entrenched giants in search and social.
How the ChatGPT advertising platform works—and why OpenAI calls it mid‑funnel
Inside ChatGPT, ads appear directly beneath unpaid responses for users on the Free and Go plans, while Plus, Pro, Business, Enterprise, and Edu accounts stay ad‑free. That placement matters: by the time an ad shows, the user is already researching, comparing products, or working through purchase‑related tasks—classic mid‑funnel behavior. OpenAI sells this format around intent, not sheer impression volume, reflecting its pitch of a shift from the “attention economy” to an “intelligence economy,” where usefulness beats interruption. Marketers can buy this inventory through partner integrations or a beta self‑serve Ads Manager offering cost‑per‑click and cost‑per‑thousand bidding plus expanded measurement tools. The platform has moved quickly since testing began in February, scaling to seven markets, with Japan and South Korea added recently and Brazil and Mexico next in line. That speed is deliberate: OpenAI needs real campaign data, not case‑study rhetoric, before it walks into an IPO roadshow talking about AI advertising revenue as a third, durable business line.
Early signals: adoption, performance, and the trust problem
OpenAI’s ad product has already crossed USD 100 million (approx. RM460 million) in annualized revenue within its first six weeks, with more than 600 advertisers signed on—serious early traction for a brand‑new format. Criteo reports more than 2,000 brands using ChatGPT ads through its integration, matching product catalogs to commercial queries inside the assistant. Yet performance and trust remain unresolved. Reported CPMs have dropped from around USD 60 (approx. RM276) in pilots to the USD 15–35 (approx. RM69–161) range as click‑through rates near 1 percent lag far behind the roughly 29 percent seen in traditional search ads. OpenAI is watching ad close‑out rates closely; the company says dismissals have fallen by 50 percent since launch, an encouraging signal that users are not rejecting the format outright. Still, several senior advertising executives argue they need stronger targeting, attribution, and measurement before shifting serious search budgets. Because the ad sits inside a tool people ask for answers, user trust is not a minor concern—it is the whole ballgame.
The $2.5B question: can AI advertising revenue reshape the market?
OpenAI’s projection of USD 2.5 billion (approx. RM11.5 billion) in ad revenue by the end of 2026 and a USD 100 billion (approx. RM460 billion) horizon by 2030 is a direct challenge to how marketers think about search and intent, even if it won’t topple the current leaders overnight. Forecasts for the entire AI chatbot ad market sit under USD 1 billion (approx. RM4.6 billion) for 2026, so OpenAI’s curve implies rapid advertiser adoption and a re‑rating of conversational search as a mainstream performance channel. At Cannes, the company’s presence in a room historically dominated by Google and Meta sent a clear signal: it wants to be considered alongside those platforms when agencies recommend experimental spend. This is also about Wall Street. OpenAI has filed confidentially for an IPO with a target valuation above USD 1 trillion (approx. RM4.6 trillion), and advertising adds a third revenue leg to subscriptions and APIs, with economics more familiar—and more appealing—to public‑market investors. The risk is that if ChatGPT’s commercial traffic remains a low single‑digit fraction of traditional search volume, the grand narrative outruns the practical reality of available budget.
What brands should do now: test, demand measurement, and watch the IPO clock
Brand marketers should treat ChatGPT ads as high‑intent, mid‑funnel inventory worth testing—but not yet worth rebalancing entire media plans around. The product is live in seven markets, with Brazil and Mexico next, and repeat spending, measurable conversions, and low user opt‑out rates will determine whether it becomes an operating business rather than a Cannes talking point. Advertisers need proof that targeting, attribution, and measurement work inside a chat product before they move larger budgets, and they should push hard on OpenAI’s trust principles: visible separation between answers and ads, no sharing of conversational data with advertisers, and long‑term user value. The privacy lawsuit over embedded analytics underscores how fragile that trust can be if data handling is opaque. The practical playbook now is straightforward: run controlled tests, compare performance against search and social, and watch how OpenAI’s IPO timeline shapes its incentive to prioritize revenue over user experience. If the company can maintain falling dismiss rates while scaling measurement credibility, ChatGPT could become the default destination for incremental, intent‑driven brand spend over the next few years.
