10x in Five Months: ChatGPT Ads Hits a $1 Billion Run Rate — While Advertisers Quietly Report 0.9% CTRs

On August 31, 2026, OpenAI announced that ChatGPT Ads — a business that did not exist in December — had reached a $1 billion annualized revenue run rate in under 200 days. On the same day, the company flipped self-service buying on across India, Europe, the Middle East, and North Africa. As of September 3, OpenAI’s help centre lists 52 countries where advertisers can now buy inventory directly through Ads Manager without ever talking to a salesperson.

The numbers are genuinely historic. When Google launched AdWords in 2000, it took years to build a billion-dollar ad business; when analysts sized up ChatGPT Ads before launch, consensus estimates landed between $500 million and $800 million for the first year. OpenAI blew through the top of that range in roughly six and a half months — faster than any advertising platform in history, according to Forbes’ coverage of the milestone.

But the milestone arrived wrapped in a paradox. The same week the company celebrated its run rate, independent analyses from Adthena and SE Ranking — and on-record agency complaints — painted a picture of a channel that still barely works as advertising. Both things are true at once, and understanding why is the key to reading OpenAI’s next phase.

From $200,000 Minimums to a Self-Service Form

The arc of the last 200 days is less about brilliance than about systematically removing friction. When ChatGPT Ads launched in the United States in January 2026, it was an enterprise product wearing a chatbot’s clothes: a reported $60 CPM, a $200,000 minimum spend, managed sales only, and a format limited to sponsored answer placements for Free-tier and $8 Go-tier users. No under-18s, no paid subscribers.

The gates came down one at a time. March brought new formats — sponsored answer cards, product spotlight units, sidebar placements — and a reported $100 million annualized run rate within six weeks. April added Australia, New Zealand, and Canada. May introduced Ads Manager and opened self-service buying in the US, which OpenAI now credits with making small and medium businesses “a material share of the business.” June brought conversion-optimized CPA bidding, the OpenAI Pixel, a Conversions API, and the UK. August 24 lit up 31 European countries. August 31 opened self-service across four more regions at once.

Along the way, the buying model inverted. In January, everything was CPM through managed sales. By August 31, OpenAI says CPC and outcome-optimized bidding account for the majority of campaigns, with Pixel and Conversions API as the measurement foundation, plus product feeds, geographic targeting, and custom audiences. Partner minimums compressed just as aggressively — from $250,000 floors to zero at StackAdapt, and from $50,000 to $10,000 at Criteo. More than 50 technology and measurement partners now plug into the stack, and inventory runs in over 40 countries even beyond the self-serve list.

The 52 self-serve markets read like a map of OpenAI’s priorities: 3 in North America, 32 across Europe including the full EEA plus the UK and Switzerland, 7 in the Gulf and Levant, 4 in North Africa, and 5 in Asia-Pacific (India, Japan, Korea, Australia, New Zealand). The absences are just as telling — no Sub-Saharan Africa, no Southeast Asia, only Brazil and Mexico in Latin America, and no China, Russia, or Turkey. Israel, notably, is not on the list; Ireland is.

A Run Rate Is Not Revenue

The fine print matters. A $1 billion annualized run rate means OpenAI took its most recent period’s ad revenue and multiplied it out to twelve months — it is a rate, not a total, and it is the standard way fast-growing ad businesses describe themselves precisely because it flatters growth. Digiday’s coverage makes the same point: the figure reflects current performance, not confirmed bookings.

Still, the slope is real: $100 million to $1 billion annualized in about five months, over a period when the addressable base went from one country to more than forty. Whatever the accounting caveats, ChatGPT Ads has crossed from pilot to platform. OpenAI frames advertising as one of four revenue pillars — alongside consumer subscriptions, enterprise offerings, and usage-based APIs — with the ad-supported free tier explicitly designed to keep ChatGPT accessible to its more than 1 billion weekly active users.

The Skeptics’ Data

Here is the paradox. An Adthena analysis covered by Campaign found a 0.91% click-through rate for one brand — roughly seven times below the 6.4% benchmark for Google Search in the same sector. One enterprise advertiser reportedly spent just 3% of a $250,000 budget over several weeks, unable to deploy the rest. Adthena executive Ashley Fletcher flagged reporting glitches in Ads Manager and concluded the platform is “still finding its feet.”

SE Ranking’s study of more than 50,000 prompts, summarized by eMarketer, went deeper: 14.35% of ads were off-topic relative to the prompt, with the share climbing above 50% in categories like news and politics. In 96.37% of placements, the advertising brand went unmentioned in the answer text — meaning the ad sits beside the answer, not inside it. And fewer than 20% of eligible users saw an ad on any given day. Agency executives told MediaPost of campaigns with ~0.6% CTRs, ~$7 clicks, and zero conversions, plus persistent discrepancies between OpenAI’s click counts and the sessions showing up in Google Analytics.

OpenAI counters with its own case studies: one e-commerce advertiser achieved a 3x return on ad spend across 28 days, and a technology partner reports that more than 80% of ad-driven ChatGPT traffic comes from new customers — a claim that, if it holds broadly, would make the channel a discovery play rather than a search-replacement play. The company says trust, relevance, and experience metrics have held up throughout the global rollout.

How to Read the Next Phase

The most important sentence in OpenAI’s announcement is the quietest one: the company plans to “explore more native ways for businesses to interact with consumers inside ChatGPT.” That is where this is going. The current format — labelled sponsored cards below answers, no influence on the answers themselves, no advertiser access to conversations — is deliberately conservative, especially in Europe, where delivery is contextual-only (conversation topic, approximate location, device, language) under GDPR legitimate interest, with the Irish DPC as supervisory authority and personalization held back behind explicit opt-in.

A 0.9% CTR on a channel with a billion weekly users is still an enormous business — and the mismatch between the $1 billion run rate and the mediocre performance data is explained by inventory growth, not efficiency: more countries, more formats, more advertisers, at prices still set by scarcity. The bet OpenAI is making is that “native ways for businesses to interact with consumers” — commerce inside the conversation, transactions the model can complete — will do for ChatGPT what the ad-loaded search results page did for Google. Advertisers paying $7 a click today are effectively funding the research.

The 200-day record will be cited for years. Whether the next 200 days produce a channel that actually performs is the question the $1 billion figure carefully does not answer.