Forty one markets, and a queue forming in Delhi

What actually opened for business#

The quiet quarry here is not the ad, it is the moment. On 27 August, OpenAI began showing ads on ChatGPT's Free and Go tiers in India, per TechCrunch and the Free Press Journal, with 50 brands at launch through agency partnerships. Dave Dugan, OpenAI's head of global ads solutions, framed it as businesses introducing themselves at relevant, high-context moments when decisions are beginning to take shape. Strip the vendor gloss and the mechanism is real: the ad arrives inside the deliberation, not beside it.

India is not a test market for ads on chatgpt. TechCrunch puts it at more than 100 million weekly active users, ChatGPT's second-largest footprint anywhere, and most of that audience sits on exactly the tiers that now carry ads. The Go plan costs 399 rupees a month, under five dollars, and was free for a promotional year. The audience was gathered first, patiently, and the stall opened second. A fox would recognise the sequencing.

The immediate question every client asks is the chatgpt ads cost question, and the honest answer has two parts: the self-serve Ads Manager opens in September with daily budgets from 725 rupees, about 7.60 dollars, per TechCrunch's reporting, while managed access runs through OpenAI's own ads portal and partners today. A floor that low is not a pricing decision, it is a recruitment drive for small advertisers, a wide gate into a young thicket.

From one market to forty one in six months#

The rollout pattern matters more than any single market. OpenAI's own help centre records the US test beginning on 9 February 2026, per the Ads in ChatGPT documentation. Eight further markets followed through the middle of the year, before 31 European countries and then India arrived in a late-August sprint that PPC Land tallies as 32 new markets in five days, 41 in total. And the pilot earned its keep fast: the February ad programme crossed 100 million dollars in annualised revenue within six weeks, per Business Today.

Notice what the shape says. Half a year of caution, then a sprint. That is a company that found its unit economics mid-summer and stopped hedging. It is also a company that confidentially filed for a stock market listing in June, per Business Today, with revenue rising to 6.7 billion dollars in the June quarter from 5.7 billion in March, while industry analysis at eMarketer, echoed across Indian trade coverage, has OpenAI anticipating 2.5 billion dollars in ad revenues this year. The ai ads business is no longer a curiosity line.

The rules OpenAI wrote for its own till#

Before any brand budget moves, read the house rules published alongside OpenAI's original testing announcement, because they are unusually specific. OpenAI's help documentation states that ads can appear below the end of a response, clearly labelled as sponsored and visually separated. Ads do not influence ChatGPT's answers, run on separate systems, and advertisers have no ability to shape, rank or alter responses. Paid tiers, Plus and above, see nothing at all.

The privacy line is drawn just as firmly, and it is the line every regulated client should quote back to their compliance team. Advertisers do not have access to chats, chat history, memories or personal details, and receive only aggregated, non-identifying performance data. Ads stay away from sensitive territory entirely: no placements near personal health, mental health or politics, and no political advertising at all, per the same documentation.

One more detail with a familiar shape: personalised ads are not initially available in the European Economic Area or Switzerland. The same regulatory weather that redrew Google's spam enforcement this weekend, which we cover in tonight's companion piece on the EEA carve-out, is shaping how the newest ad platform ships its targeting. Borders are becoming product features, and any chatgpt marketing plan that ignores jurisdiction is already stale.

Weird is one word for it. Unfamiliar is fairer. The auction logic, per OpenAI, weighs expected relevance and outcomes using the conversation's context and intent, the ad's landing page and copy, and, where the user allows it, personalisation signals. That is closer to quality-score thinking than to social-feed thinking, which is why search-trained teams will adapt to ai ads faster than social-native ones. The scent is search intent, whatever the burrow it surfaces in.

Running ads on ChatGPT without fumbling it#

The question circling every marketing den this week, why does chatgpt now have ads?, has a plain answer: the infrastructure bill and the listing. The playbook for ads on chatgpt writes itself if you have run launch-window media before. Early inventory in a new auction is systematically underpriced because demand has not arrived yet, and a 725 rupee floor makes testing nearly free by paid-search standards. The scent trail here is short and fresh: sign up, seed conservative campaigns, learn the reporting surface before your competitors clear their September budget meetings.

Position ads on chatgpt honestly inside the mix. This inventory rewards the same assets that win paid search: tight landing pages, specific claims, clean measurement. It punishes the same laziness paid social punishes: creative that interrupts rather than answers. And for categories OpenAI gates as sensitive, eligibility criteria decide entry, so fintech and healthcare advertisers should read the policy before promising anyone a launch date.

Our own coverage has tracked this platform since the US test: how budgets behave in the nine-week growing pains, what the first international expansion signalled, and where conversational placements fit for specific verticals. India is the largest single test of whether answer-engine advertising scales beyond rich-market novelty, and the evidence starts arriving the week the self-serve till opens.

Frequently asked questions#