In February 2026, OpenAI quietly began testing ads inside ChatGPT with a small group of enterprise advertisers who had to commit a minimum of $50,000 to participate. By May 5, 2026, that floor was gone. OpenAI launched ChatGPT Ads Manager — a self-serve platform that lets any US advertiser create, manage, and measure conversational ad campaigns without a sales rep or a six-figure budget.
The announcement is significant beyond the platform itself. ChatGPT is the first AI assistant at serious scale to build a native advertising layer, and the model it chooses will influence how AI advertising works across the industry.
This review covers how the platform works, what the ad unit looks like, what advertisers pay, and whether the product is ready for performance marketing investment in mid-2026.
What ChatGPT Ads Manager Is
ChatGPT Ads Manager is OpenAI’s self-serve ad buying platform. It is accessible at OpenAI.com and allows advertisers to:
- Register as an advertiser and add payment information
- Set campaign budgets, bids, pacing, and country targeting
- Upload ad creative and launch campaigns
- View performance data in a reporting dashboard
It supports two buying models: cost-per-mille (CPM) impressions and cost-per-click (CPC) clicks. CPC bidding was added May 5, 2026, when the self-serve beta launched. Before that, only managed CPM buys were available, with a $50,000 minimum spend.
Campaign structure mirrors Google Ads loosely: budgets and dates are set at the campaign level; bids and context hints are set at the ad-group level.
The Ad Unit: chat_card
ChatGPT uses a single primary ad format internally called the chat_card.
The unit appears below the AI response, never inside it. It is clearly labelled as sponsored. The card contains:
- Advertiser name and favicon
- Headline
- Short description (body copy)
- Image (optional)
- Destination URL
For e-commerce advertisers who connect a product catalog, the unit can render one or more individual products from that catalog inside the same sponsored card.
OpenAI’s ad policy is explicit on one point: ads cannot influence ChatGPT’s answers. The assistant generates its response independently; the ad unit is rendered separately, after the answer, based on contextual signals from the conversation. If the assistant recommends a competitor product, the sponsored card below may show your product — but the editorial recommendation is not for sale.
Pricing
CPC: $3–$5 per click
Cost-per-click bids currently range from $3 to $5, based on screenshots and trade reporting verified at launch. This puts ChatGPT CPC roughly in line with mid-funnel Google Search bids for moderately competitive queries, and meaningfully above the average display CPC.
CPM: $25–$60 per thousand impressions
CPM pricing ranges from $25 to $60 depending on targeting parameters and inventory quality. The $60 CPM that OpenAI commanded at the managed pilot launch in February eroded to as low as $25 within ten weeks as more inventory became available. Supply is growing faster than demand at this stage.
Minimum spend: None
The previous $50,000 pilot threshold was removed with the self-serve launch. There is no stated minimum spend for the self-serve beta, making the platform accessible to small and mid-market advertisers for the first time.
Revenue Context
OpenAI’s advertising ambitions are large. The company is targeting:
- $2.5 billion in ad revenue in 2026
- $100 billion in ad revenue by 2030
The $100 billion figure would make OpenAI’s ad business comparable in scale to Meta’s current advertising operation. That projection depends heavily on ChatGPT maintaining and growing its user base, expanding internationally, and proving that conversational ad formats can generate conversion rates that justify the current CPC and CPM pricing.
ChatGPT has approximately 800 million monthly active users as of early 2026. Monetizing even 20% of that audience at scale would make the ad revenue targets plausible — but the targeting capabilities, measurement tools, and advertiser trust needed to support those numbers are still being built.
Geographic Availability
The self-serve Ads Manager launched in the United States only. OpenAI has announced plans to expand the ad pilot to:
- United Kingdom
- Mexico
- Brazil
- Japan
- South Korea
Timing for international rollout was not specified beyond “coming months.”
What Advertisers Should Know
The format is honest but limited
The chat_card unit is genuinely separated from the AI answer. Users can see clearly that it is sponsored. That transparency is a strength — but it also means advertisers cannot benefit from editorial halo. The ad is adjacent to the response, not inside it.
Targeting transparency is limited
Unlike Google Ads or Meta Ads, ChatGPT Ads Manager does not expose granular audience targeting controls in its beta form. Context hints can be set at the ad-group level, but the exact signals used to match an ad to a conversation are not publicly documented. Advertisers are buying into a contextual matching system they cannot fully inspect.
There are no established conversion benchmarks
ChatGPT advertising is new. No large-scale, independent studies exist comparing ChatGPT CPC conversion rates to Google Search CPC conversion rates for equivalent query types. Early adopters are building the benchmark data. Whether a $3–$5 CPC in a conversational AI context produces the same downstream conversion rate as a $3–$5 Google Search CPC is the core open question.
The no-influence policy is meaningful
OpenAI has staked a clear position: the assistant’s answers are not for sale. This protects advertiser credibility in the medium term — users who trust ChatGPT will not feel they are reading a paid-influenced recommendation. But it also limits what conversational advertising can do. If a user asks ChatGPT to recommend the best project management tool and the assistant recommends a competitor, your sponsored card below the answer is working against the editorial content it is adjacent to.
Competitive Context
ChatGPT Ads Manager has no direct competitors today. Google has not yet launched a comparable self-serve ad layer inside Gemini, though Google’s dominance in AI search (Gemini 3.5 Flash now powers AI Overviews globally) makes that a near-certainty. Perplexity runs sponsored follow-up questions, a different format at a much smaller scale.
OpenAI has a window to define what AI-native advertising looks like before the major platforms catch up. The chat_card format, the no-influence policy, and the CPC pricing model are all early choices that will either become industry standards or be revised as the data comes in.
Rating: 3.5 / 5
ChatGPT Ads Manager is an honest first version. The ad unit is clean, the no-influence policy is credible, and removing the $50K entry barrier opens the platform to the long tail of advertisers who will actually build the case studies needed to prove conversational AI advertising works.
The limitations are real: targeting transparency is thin, conversion benchmarks don’t exist yet, and the $2.5B revenue target implies a pace of development that will require both the ad product and the user base to scale simultaneously.
For performance marketers: worth testing a modest budget now to build internal baseline data before competition drives CPC prices up. For brand advertisers: the CPM erosion from $60 to $25 in ten weeks suggests pricing will be favorable for awareness campaigns through at least mid-2026.
For everyone else: the most interesting question is not whether OpenAI reaches $2.5B in ad revenue this year, but whether the model — ads that don’t influence answers, placed adjacently, in a conversational context — proves durable user trust. If it does, every AI assistant will follow. If it doesn’t, the format will be quietly revised.
ChatForest covers AI tools for professionals and developers. This review is based on publicly available documentation, trade reporting, and OpenAI’s published ad policies — not direct advertiser access to the platform.