ChatGPT has plenty of advertisers, but not much revenue
OpenAI has expanded access, lowered costs, and added targeting tools, but advertiser growth hasn’t yet translated into enough revenue.
Spy on Any Website
OpenAI has plenty of advertisers for ChatGPT, but that has yet to translate into significant ad revenue.
The number of ChatGPT advertisers rose from about 20,000 at the end of August to roughly 60,000 in early October, according to Dan Salmon, a partner at New Street Research. Clicks cost about $5, compared with $30 to $50 for similar commercial software ads on Google. Even with that growth, OpenAI is still likely to miss the $2.4 billion in advertising revenue it had projected to investors for 2026, according to The Information.
That’s a useful snapshot of where ChatGPT advertising sits right now. OpenAI has lowered the barriers to trying it, expanded into more than 60 countries, added self-service buying, and started building the targeting and measurement tools performance marketers expect.
Marketers appear willing to experiment. The harder job is turning that experimentation into sustained budgets.
ChatGPT ads are starting to look like an ad platform
The early version of ChatGPT advertising looked more like a tightly controlled pilot. OpenAI initially charged on a CPM basis, required a Super Bowl-like $200,000 minimum commitment, and limited access to a small group of large advertisers.
That model changed quickly.
OpenAI now supports CPM and CPC bidding, with recommended CPC bids of $3 to $5. It has also tested CPA bidding with selected advertisers, though OpenAI has not publicly confirmed that pricing model.
Self-service buying is also spreading. When OpenAI expanded ChatGPT Ads into Indonesia, Malaysia, the Philippines, Singapore, Thailand, Vietnam, and Taiwan in September, businesses in those markets could buy through Ads Manager from launch. The expansion pushed availability past 60 countries.
The world’s most powerful SEO platform, purpose-built for Enterprise.
That changes the practical calculation for advertisers. A $200,000 pilot belongs in a planning meeting. A $5 click belongs in a test budget. It also says advertisers aren’t breaking down the doors to get on board.
All of which casts further doubt on OpenAI’s claim that less than 200 days after launch, ChatGPT Ads reached a $1 billion annualized revenue run rate. The company offered no proof of this, and, because it is privately owned, there is no way to confirm it is true. Furthermore, ARR is not a legally accepted accounting method. It is an extrapolation of a current growth rate — whether that’s per day, per week, or per month — over 12 months. Both OpenAI and its rival Anthropic use the measure frequently.
The targeting starts with the conversation
ChatGPT gives OpenAI a targeting signal that traditional ad platforms don’t have in quite the same form: the conversation that happens immediately before the ad appears.
The company says ad selection can consider the context and intent of the current conversation, along with advertiser-supplied descriptions of relevant topics, the ad and landing page, and other signals when personalization is enabled. Advertisers can provide “context hints” describing conversations or topics where an ad might be relevant, rather than relying on exact-match keywords.
Advertisers can also upload email addresses and phone numbers to match customer lists. According to Gadget Review’s account of OpenAI’s documentation, those matched audiences require at least 25,000 users.
For marketers, that puts ChatGPT somewhere between search advertising and audience advertising. The current conversation can reveal immediate intent, while customer matching adds the familiar first-party-data layer.
It also creates a different brand-safety problem. An ad isn’t appearing next to a webpage whose subject can be classified in advance. It can appear around a private conversation whose direction changes message by message.
OpenAI says it is working with DoubleVerify and Integral Ad Science on brand-suitability evaluation and has safeguards intended to keep ads away from sensitive or emotionally vulnerable conversations. The company also says those partners don’t receive access to private conversations when evaluating those protections.
OpenAI is pushing beyond the standard ad click
The more interesting experiments start after someone sees the ad.
OpenAI is testing Sponsored Agents in the U.S., which let a person click an ad and enter a separate conversation with an advertiser’s AI agent. In OpenAI’s example, someone shopping for a dining table can ask the advertiser’s agent whether it will fit a room, continue asking questions, and then follow a link to the company’s site.
That changes the role of the ad. Instead of sending every click immediately to a landing page, the advertiser can put another conversational layer between interest and the website.
For marketing teams, that creates work beyond media buying. Someone has to decide what the branded agent knows, which claims it can make, how current its product information is, when it hands someone off, and how those interactions get measured.
OpenAI has also connected Ads Manager with HubSpot and Shopify and added tools to create or modify campaigns via natural-language prompts. An optional feature can adapt an advertiser’s existing headlines and descriptions to the conversational context.
Those features move ChatGPT advertising closer to the rest of the martech stack. They also increase the number of places where campaign governance can break.
Visual ads widen the inventory
OpenAI is also moving beyond text placements.
This month, the company announced a U.S. test of visual ads shown alongside image-generation results. OpenAI says the ads will remain clearly labeled and separate from the generated image, and won’t influence ChatGPT’s answer.
The format gives advertisers another context to work with. Someone generating an image may be exploring a product, design, trip, event, or other idea where a visual placement makes more sense than a text ad.
OpenAI is building out measurement at the same time. Its partners now include Hightouch, Tealium, and LiveRamp for conversion data, as well as attribution and measurement providers such as AppsFlyer, Adjust, Branch, Triple Whale, Kochava, Haus, Measured, and others.
That work is important because advertiser growth by itself tells marketers very little about performance. Sixty thousand advertisers can mean strong repeat demand, or it can mean 60,000 companies running inexpensive experiments.
The reported revenue shortfall makes that distinction worth watching. OpenAI has made ChatGPT ads cheap enough and accessible enough to attract plenty of tests.
The next test is harder: whether marketers keep the channel in the budget after they have enough conversion, incrementality, and customer-quality data to compare it with everything else they already buy.
MarTech is owned by Semrush. We remain committed to providing high-quality coverage of marketing topics. Unless otherwise noted, this page’s content was written by either an employee or a paid contractor of Semrush Inc.
Add us as a preferred source on Google
Google's "preferred sources" feature allows users to customize their search results by selecting news outlets they want to see more often in the "Top Stories" section.
Add Martech Now