OpenAI’s ad pilot just crossed $100 million in annualized revenue in under six weeks. More than 600 advertisers are in. The company is expanding to Canada, Australia, and New Zealand within weeks, with “many more markets” later this year.
If you stopped reading there, you’d think ChatGPT ads are the next Google Ads. The reality is messier, and every agency owner and marketing director should understand why before reallocating budget.
The numbers behind the headline paint a picture of a platform still figuring out the basics: click-through rates roughly seven times lower than Google Search, a $60 CPM that prices out most SMBs, broken reporting tools, and an audience ceiling that excludes every paying ChatGPT subscriber. For agencies managing client budgets, this matters more than the revenue milestone.
What $100 Million in Revenue Actually Means
The $100M figure is annualized recurring revenue, not actual collected revenue. OpenAI hit that run rate by selling high-CPM placements to a select group of enterprise advertisers with $200,000 minimum commitments. Brands like Target, Adobe, Williams-Sonoma, and Albertsons are in the pilot.
That’s impressive for six weeks. It’s also a very specific slice of the market.
When CNBC reported on advertiser frustration in mid-March, one enterprise advertiser had spent just 3% of a $250,000 budget over several weeks. The money was committed to the trial, couldn’t be deployed elsewhere during the quarter, and the advertiser couldn’t even see whether any of those impressions drove meaningful action.
That’s not a performance advertising channel. That’s a brand awareness experiment with an enterprise price tag.
The Click-Through Rate Problem
According to Adthena research, ChatGPT ads are generating a 0.91% click-through rate. Google Search averages 6.4%. That’s not a rounding error. It’s a structural gap that reflects how people use ChatGPT versus how they use a search engine.
Search users are browsing. They scan results, click through options, explore. ChatGPT users are completing tasks. They ask a question, get an answer, and move on. An ad sitting at the bottom of a completed answer is competing with the user’s instinct to close the tab.
WIRED’s experiment asking ChatGPT 500 questions found that roughly one in five prompts triggered an ad. The ads were contextually relevant, covering everything from dog food to MBA programs to cruise vacations. But relevance alone doesn’t solve the behavioral mismatch.
This isn’t a problem OpenAI can fix with better targeting. The format itself fights the user’s intent. When someone asks ChatGPT “what’s the best CRM for a 10-person sales team” and gets a thorough answer, they’re less likely to click a Salesforce ad at the bottom than they would be to click a Salesforce result on Google, where they’re actively shopping.
$60 CPM: Three Times Meta, With No Attribution
At $60 CPM, ChatGPT ads cost roughly three times what Meta charges. The difference: Meta gives you granular conversion tracking, audience segmentation, lookalike modeling, and a decade of optimization data. ChatGPT gives you impressions.
OpenAI’s Ad Manager tool, the reporting interface for advertisers, has been plagued by a technical glitch that prevents advertisers from viewing their own campaign performance data. You’re paying a premium CPM and you can’t see whether the spend did anything.
Some advertisers have resorted to workarounds: UTM-tagged URLs in ad creative, tracking lifts in branded search volume, monitoring direct traffic spikes during campaign periods. These are the kinds of measurement hacks you use when a platform doesn’t have real attribution, not when you’re paying $60 per thousand impressions.
For agencies managing SMB clients with $5,000 to $20,000 monthly ad budgets, the math doesn’t work. A $200,000 minimum spend buys approximately 3.3 million impressions at $60 CPM. At a 0.91% CTR, that’s roughly 30,000 clicks, or about $6.67 per click, with no way to measure what happens after.
Compare that to Google Ads or Meta campaigns where you can track cost-per-lead, cost-per-acquisition, and ROAS down to the keyword level.
The Audience Ceiling Nobody’s Talking About
Here’s the structural problem that should give every performance marketer pause: ChatGPT ads only reach free and Go ($8/month) users. Every paid subscriber (Plus at $20/month, Pro at $200/month, Business, Enterprise) never sees an ad. OpenAI has positioned ad-free access as a permanent feature of paid tiers, not a temporary restriction.
That means advertisers are buying reach among OpenAI’s least committed users. The power users, the ones who depend on ChatGPT enough to pay for it, the ones making purchasing decisions and researching business tools, are completely excluded.
Reuters reported that roughly 85% of free and Go users in the U.S. are eligible to see ads, but less than 20% are shown them on any given day. OpenAI is deliberately throttling ad frequency, which is good for user experience but limits the scale that advertisers need for statistical significance in their tests.
Google and Meta don’t have this problem. Their highest-value users see the same ads as everyone else. On ChatGPT, the audience with the most purchasing intent and the highest lifetime value is permanently behind a paywall.
For B2B advertisers, this is especially problematic. Decision-makers at companies using ChatGPT for work are overwhelmingly on paid plans. The free tier skews toward casual, personal use. If you’re selling enterprise software, professional services, or anything with a considered purchase cycle, you’re advertising to the wrong audience by design.
What This Means for Agencies Right Now
If you’re an agency owner or marketing director being asked by clients whether they should test ChatGPT ads, here’s the honest assessment:
Not yet, for most. The minimum spend is too high, the measurement is too weak, and the audience targeting is too limited for anyone who needs to show ROI on their ad spend.
Maybe, for large brands with experimental budgets. If you have a client with $200K+ to allocate toward brand awareness in a new channel, and they understand it’s a test-and-learn exercise rather than a performance campaign, there’s value in getting early data.
Watch closely, for everyone. OpenAI hired former Meta advertising executive David Dugan to run ad sales. Criteo is integrated as the first ad tech partner. Impression volume jumped approximately 600% from early to mid-March. The infrastructure is being built. When measurement catches up, the early movers will have an advantage.
The real question isn’t whether ChatGPT ads will work eventually. OpenAI has 400 million weekly users and a product people return to daily. The question is when the platform will mature enough to justify performance budgets, and what the ad format will look like when it does.
The Bigger Play: Getting Cited, Not Just Getting Seen
Here’s what most coverage of ChatGPT ads misses: while brands are paying $60 CPM to appear below ChatGPT’s answers, the brands being mentioned inside those answers pay nothing.
When someone asks ChatGPT which marketing agencies specialize in healthcare AEO and your agency gets cited in the response, that’s a recommendation with more weight than any banner ad. You’re not interrupting the user’s experience. You’re part of it.
This is the difference between paid placement and earned authority in AI search. One costs $60 per thousand impressions. The other costs the effort of building content that AI models trust enough to cite.
We’ve seen this play out directly. Seasons in Malibu holds 4,200+ keyword rankings, 814K+ monthly social impressions, and averages 5 patient admits per month driven directly through Emarketed’s marketing. A significant part of that visibility now comes from AI citations, not paid placement. They’re ranked #1 on Perplexity for core treatment queries, a position that delivers referrals without a media budget attached.
The agencies that will win this transition aren’t the ones racing to test ChatGPT ads at $60 CPM. They’re the ones building the content authority and citation infrastructure that makes their clients the answer, whether or not ads are involved.
What to Do This Week
If you manage paid media budgets: Brief your clients on where ChatGPT ads actually stand. Set expectations. The $100M headline is a revenue milestone for OpenAI, not a signal that agencies should redirect spend from proven channels.
If you manage organic and content strategy: Double down on AEO and citation optimization. Every day that competitors focus on paid AI placements while you build earned AI visibility is a day you pull further ahead.
If you manage both: Start tracking which client queries trigger ChatGPT ads, and which trigger organic citations. That data will tell you where the value actually sits for each client’s category.
The paid side of AI search will mature. OpenAI is too well-funded and too motivated for it not to. But right now, in March 2026, the gap between what ChatGPT ads promise and what they deliver is wide enough that any agency spending client money there should be doing it with eyes open, not because a $100M headline made it sound like a sure thing.
FAQ
Are ChatGPT ads worth testing for small businesses? Not yet. The $200,000 minimum spend and $60 CPM pricing put ChatGPT ads out of reach for most small businesses. Even if the minimum drops, the 0.91% click-through rate and lack of conversion tracking make it impossible to measure ROI, which is exactly what small businesses need before committing budget.
How do ChatGPT ads compare to Google Ads performance? ChatGPT ads run at a 0.91% CTR versus Google Search’s 6.4% average. Google also offers mature attribution, conversion tracking, and audience segmentation that ChatGPT currently lacks. At $60 CPM (roughly 3x Meta’s rates), ChatGPT is asking for a premium without delivering premium measurement.
Will ChatGPT ads get better over time? Almost certainly. OpenAI has hired senior Meta advertising talent, integrated Criteo as an ad tech partner, and is expanding to international markets. But “better over time” doesn’t help agencies that need to justify spend this quarter. Wait for measurement tools, lower minimums, and enough data to benchmark performance before making it a line item.
What’s the difference between ChatGPT ads and AI citations? ChatGPT ads are paid placements that appear below the chatbot’s answers. AI citations are organic mentions within the answer itself. Citations carry more trust (the AI is recommending you, not displaying your ad) and cost nothing beyond the content investment. For most businesses, building citation authority through answer engine optimization delivers more sustainable results than paying for ad impressions in a platform that’s still in pilot.
Which brands are advertising on ChatGPT? The pilot includes enterprise brands like Target, Adobe, Williams-Sonoma, and Albertsons, along with ad placements from agencies including Omnicom, WPP, and Dentsu. More than 600 advertisers are currently participating, and Criteo is serving as the first ad tech integration partner.
Should agencies recommend ChatGPT ads to clients? Only for clients with large experimental budgets and no immediate ROI requirements. For performance-focused clients, the combination of high CPM, low CTR, broken reporting, and a structural audience ceiling (no ads shown to paid subscribers) makes it a poor fit today. Track the platform’s development and revisit when measurement matures.
