OpenAI’s experiment with advertising inside ChatGPT is moving a lot faster than most people expected.
Just six weeks after launching a U.S. pilot, OpenAI says its ChatGPT ads business has already crossed $100 million in annualized revenue — a milestone that signals two things at once: advertiser demand is real, and OpenAI is actively building a second revenue engine beyond subscriptions.
What’s actually being tested
OpenAI began showing ads in January to some U.S. users on the Free tier and the Go plan. The company says these ads are separate from ChatGPT’s generated answers, do not influence responses, and user conversations aren’t shared with marketers.
That separation is the key promise: ads exist, but they aren’t supposed to contaminate the assistant’s output.
The “room to grow” stat is the tell
Here’s the number that explains why investors and advertisers are paying attention:
- About 85% of eligible users could see ads
- But fewer than 20% are actually shown ads on a daily basis
That gap means OpenAI can increase monetization without even adding new users — simply by expanding frequency, placement, or eligible surfaces.
Advertisers are already lining up
OpenAI says the pilot has expanded to 600+ advertisers, and nearly 80% of small and mid-sized businesses it surveyed signaled interest in running ChatGPT ads.
That’s a strong early signal that brands view ChatGPT as a new kind of “intent environment” — closer to search behavior than social scrolling.
Next step: self-serve ads in April
OpenAI plans to launch self-serve advertiser capabilities in April, which would open the door to a much larger pool of advertisers and faster scaling (similar to how Google and Meta grew).
It has also hired David Dugan, a former Meta ads executive, to lead its global advertising solutions team — another sign this isn’t a casual experiment.
International expansion is coming (including Canada)
OpenAI says it plans to expand the ad test to more countries in the coming weeks, including Australia, New Zealand, and Canada.
The balancing act: revenue vs. trust
OpenAI claims it’s seeing no negative impact on trust metrics, with low ad dismissal rates and improving relevance as it learns from feedback.
But the risk is obvious: ChatGPT isn’t Instagram. People use it to think, decide, and solve problems. Even if ads don’t alter answers, the mere presence of ads can make users wonder what’s being optimized — and skepticism can spread fast.
Bottom line: OpenAI’s ad pilot clearing a $100M run-rate this quickly suggests advertising could become a major pillar of the company’s future — especially as AI infrastructure costs keep climbing. The real test won’t be revenue. It’ll be whether OpenAI can scale ads without cracking the one thing ChatGPT depends on: trust.
