OpenAI Is Selling Ads Inside the Answer. That Changes the Math on Everything.
OpenAI described its ad pilot as a way to help users "discover useful products and services." I've read that sentence three times now and I still can't figure out who it's supposed to convince. The more interesting version: the company generating your answer is now selling ad space next to that answer, and it crossed $100 million in annualized ad revenue in under two months. That's not a test. That's a business.
And for anyone running paid media, this creates a problem that doesn't have a clean precedent. Google sells ads above organic results, sure. But Google doesn't write the organic results. ChatGPT does both. It generates the answer and then sells the real estate around it. That structural conflict is going to matter more than most of the coverage is acknowledging right now.
The speed tells you more than the press release
$100M ARR in under 60 days. CNBC reported that number last week, and it's worth sitting with for a second. For context, Meta's Threads ads pilot hasn't disclosed comparable early figures. TikTok took considerably longer to reach that pace in its early US ad rollout. OpenAI got there with 600-something advertisers, a $200K minimum entry bar, and reportedly low CTRs.
That math only works if brands are paying a serious premium per impression. Which they are, because the targeting is conversational context. If someone asks ChatGPT "what's the best mid-size SUV for a family of four," and a Ford or Mazda ad appears alongside the response, that's about as high-intent as it gets. In one account we manage (roughly $120K/month across search and social), the team has been tracking ChatGPT referral traffic since January. It's still small, maybe 2-3% of organic, but it's growing at about 15% month over month. The trajectory is the thing that keeps coming up in our planning calls.
Truist's estimate is under $1B in ad revenue for 2026, growing to $30B by 2030. I'm somewhat skeptical of the 2030 number because analyst projections four years out are basically horoscopes. But the near-term trajectory seems right. This thing is accelerating.
The conflict nobody wants to name out loud
Here's where it gets uncomfortable. OpenAI says ads don't influence ChatGPT's answers. They say conversations are kept private from advertisers. They say there's been "no impact on privacy-related trust metrics."
On paper, that sounds like a clean separation. And maybe it is, today, with 600 advertisers and a managed pilot. But the incentive structure is pointed in one direction. If ad revenue scales to even $5B, there will be enormous pressure to make the ad product more performant. More performant means better targeting. Better targeting means using more of the conversation. I don't think OpenAI is lying about the current setup. I think the current setup is temporary, and the pressure to evolve it is going to be relentless.
Think about it this way. Google started with "Don't be evil" and ten blue links with clearly labeled ads above them. Twenty years later, the top four results on most commercial queries are ads, and the organic results got pushed below the fold on mobile. That didn't happen because of malice. It happened because the ad business grew and the incentives followed. I'd expect a similar drift here, probably faster.
Who's already in and what they're buying
The advertiser list is revealing. Target, Adobe, Audible, Williams-Sonoma, Ford, Mazda. Agency holding companies WPP, Omnicom, and Dentsu all secured placements. Criteo joined the pilot too, which is the commerce media angle. These aren't experimental budgets from innovation teams. These are established media buyers placing real money.
The $200K minimum tells you something about who this is for right now. It's enterprise. It's the brands that already have agency relationships and can eat a test budget without blinking. Self-serve is coming. OpenAI is reportedly testing a ChatGPT Ads Manager. When that launches and the minimum drops, expect the competitive dynamics to shift fast. The early movers are buying exclusivity as much as impressions.
The geographic expansion is also moving quickly. US first, now Canada, Australia, New Zealand, with "many more markets" planned for later this year. If you run international campaigns, the arbitrage window in new markets is probably going to be real. Early CPMs in fresh geos tend to be meaningfully cheaper before local competition ramps up.
The organic traffic problem just got worse
This is the part that keeps nagging at me. If you've been watching your organic search traffic erode as AI Overviews eat into click-through rates, ChatGPT ads create a particularly frustrating dynamic. The AI answers the question directly, which means fewer people click through to your site. And now, to get visibility inside that AI-generated answer, you might need to buy it back.
It's a bit like watching your parking spot get converted into a paid lot. You used to park there for free. Now someone else owns the spot and is charging rent.
For SEO-dependent businesses, this is a second front opening up. Google's AI Overviews were the first. ChatGPT ads are the second. And unlike Google, where you can at least optimize content to appear in the AI Overview, ChatGPT's ad placement is purely a paid product. You can't earn your way in. We wrote about how experimental ad budgets are growing not because brands want to spend more but because they're being forced to. ChatGPT ads are another example of that same pressure.
The CTR problem and why it might not matter
Some insiders have told CNBC they're frustrated with the slow rollout and that CTRs are low compared to search ads. That's probably true and also probably beside the point. The comparison to search CTRs is misleading because the user behavior is fundamentally different. Someone searching Google "best CRM for small business" and clicking an ad is in buying mode. Someone asking ChatGPT the same question might be in research mode, learning mode, comparison mode. The intent is less transactional, which means the CTR will naturally be lower.
But the signal quality could end up being higher. If someone is having a multi-turn conversation with ChatGPT about their business needs and then sees a relevant ad, the context around that impression is extraordinarily rich. I think the brands that figure out how to value that context correctly, rather than just comparing raw CTR to Google Search, are going to find something useful in this data. From what I've seen in other new-platform launches, the first 6 months of performance data is almost always misleading because nobody has figured out the right creative format or the right measurement framework yet.
Whether to test, and how to think about the budget
If you have $200K to commit and you're in a category where conversational intent matters (SaaS, consumer electronics, financial services, auto, education), I'd probably test it. Not because the current returns will justify the spend. They likely won't. But because the data you collect now, what queries trigger your ads, what the conversation context looks like, what the early engagement patterns are, that data becomes genuinely valuable when self-serve opens up and the platform matures.
If you don't have $200K, wait for self-serve. It's coming. When it does, start small. I'd allocate maybe 5-8% of your experimental budget (not your core budget) and run it for 30 days alongside your existing search campaigns. Measure branded search lift and direct traffic changes, not just in-platform CTR. The attribution on these early AI ad platforms is going to be messy for a while, and honestly, pretty inconsistent too.
One thing I wouldn't do: ignore it entirely. The speed of adoption here, both from OpenAI's revenue side and from the advertiser side, suggests this isn't a novelty. The holding companies don't commit resources to novelties. They commit to channels they expect to bill against for years.
The question that should keep paid media teams up at night
The real issue isn't whether ChatGPT ads work today. It's what happens when the platform that generates the answer also controls the ad inventory, and both of those things are optimized by the same AI. Google at least had a wall, however thin, between organic ranking and ad auction. ChatGPT doesn't have that architecture. The answer and the ad come from the same system.
I don't think OpenAI is going to do anything overtly manipulative. But I also don't think the current separation will survive contact with a $30B revenue target. And to be fair, this isn't entirely new. Meta has always shaped what people see in the feed while also selling placements in that same feed. It just feels a lot more direct when the AI is literally writing the answer your customer reads.
The most honest framing I can offer: if you're a paid media buyer, you're probably going to end up spending money here within 18 months whether you want to or not. The only real choice is whether you start learning the platform dynamics now, while the data is cheap, or later, when everyone else has already figured it out and the CPMs reflect that.
By Notice Me Senpai Editorial
