OpenAI Is Selling Ads Inside ChatGPT — Targeting by Prompt, Not by Person
A leaked StackAdapt deck reveals OpenAI is running a limited ChatGPT advertising pilot with prompt-relevance targeting, floor CPMs of $60, and a $50,000 minimum spend. Unlike traditional programmatic, the system has no audience targeting or demographic overlays — ads appear based on what users are asking, not who they are. OpenAI is targeting $2.5B in ad revenue this year.
Original sourceA leaked internal deck from StackAdapt, an independent demand-side platform, reveals that OpenAI is actively recruiting advertisers for a limited ChatGPT advertising pilot — and the mechanics are unlike anything in the current programmatic ecosystem.
The system eliminates the standard levers of digital advertising: no audience targeting, no demographic overlays, limited geographic controls. Instead, ads are served based on "prompt relevance" — matching commercial intent signals in what users are actively asking to relevant advertisers. The architecture is closer to sponsored search than display advertising, but operating inside a conversational interface that has no obvious ad placement.
Pricing reflects the experimental premium: the minimum bid is $60 CPM, with floor pricing at $15 for single-advertiser niche placements. Entry requires a $50,000 minimum spend commitment through end of May. OpenAI clarified to Adweek that the $15 floor is not selectable by advertisers — $60 is the actual minimum bid, though delivery may clear at lower effective rates in a "proto-auction model."
The strategic context is stark: OpenAI is targeting $2.5 billion in advertising revenue in 2026, with ambitions of $100 billion by 2030. For a company spending aggressively on compute while racing to profitability, ads represent the fastest path to non-subscription revenue at scale. The question is whether users will tolerate commercial content appearing in responses they've come to trust as neutral.
For the AI industry broadly, ChatGPT ads would mark the moment the conversational AI interface becomes a mainstream advertising medium — with implications for every competitor building chat-first products. If prompt-relevance targeting proves effective, it could displace significant search advertising spend and establish a new category of intent-based commercial targeting that makes keyword search look crude by comparison.
Panel Takes
The Builder
Developer Perspective
“Prompt-relevance targeting is actually a more sophisticated signal than keywords — if a user asks 'what's the best project management tool for a 10-person engineering team,' that's higher commercial intent than a search query ever conveyed. The $60 CPM floor is steep but rational for intent quality this high. Worth watching conversion data carefully.”
The Skeptic
Reality Check
“Injecting commercial content into a medium users trust as neutral advice is the fastest way to destroy that trust. Every response that might be commercially influenced becomes suspect, and the $50,000 minimum means only large brands play — making the bias systematic and predictable. OpenAI is betting users won't notice or won't care. That's a dangerous bet.”
The Futurist
Big Picture
“The advertising layer landing on conversational AI marks a civilizational shift in how commercial intent gets mediated. The entire $600B search advertising industry was built on keyword matching — prompt-relevance targeting operating at the semantic level of actual user goals is categorically more powerful. This is either the foundation of OpenAI's path to profitability or the moment it begins losing the trust that made it dominant.”