Who can advertise in ChatGPT? Industries, countries and the political question

The short answer

Eligibility for ChatGPT advertising is decided by two lists, and they are different lists. The first is where your business is registered: advertiser accounts opened in the United States and a small group of markets first, and access has widened through 2026 to include Canada, the United Kingdom, Australia, New Zealand, Japan, South Korea, Brazil and Mexico, with European markets reported joining in late August 2026. The second is what you sell: lightly regulated categories are routinely approved, while finance, healthcare, legal, gambling and political advertising are restricted or reviewed case by case.

There is a third gate people forget: who actually sees the ads. Placements serve to users on the free and Go tiers only, in the countries where serving is live. Whoever you are, you are not buying the whole ChatGPT audience — you are buying the part of it that does not pay.

Everything above is a snapshot of a market that changes month to month, which is rather the point of this guide: the durable knowledge is how the gates work, not today’s list. We confirm the current lists per brief as part of the managed ChatGPT advertising service.

Gate one: the country lists

The platform distinguishes between the countries whose businesses can open an advertiser account and the countries whose users are served ads. They overlap but are not identical, and each has expanded on its own schedule through 2026. The business country and currency lock at account setup, so a multi-market advertiser needs to choose its registration deliberately rather than defaulting to whichever entity signed up first.

For advertisers outside the self-serve countries, the practical route is programmatic: demand-side platforms holding early access to AI-assistant inventory can reach the same placements from outside the self-serve programme, bought alongside display, video, connected TV and audio. That route is compared properly in how to buy ads in ChatGPT.

Gate two: the category rules

The pattern across every early report is consistent: the platform approves the categories that carry the least regulatory weight and slows or blocks the rest.

Routinely approved: retail and e-commerce, consumer goods, travel, education, digital products, and general B2B products and services.

Restricted or manually reviewed: finance and financial products, healthcare and health claims, legal services, gambling, alcohol and tobacco, dating, and political advertising. “Restricted” spans a range — some categories face extra review, some are effectively unavailable — and the boundary moves as the platform matures. What does not move is the posture: if your claims are regulated in the offline world, expect the platform to treat them cautiously here.

The declared activity category shapes the account from setup, and landing pages must match what the ad promises. Neither is a formality; both are checked.

Gate three: the audience tiers

Ads render for free-tier and Go-tier users only. Subscribers on the higher paid plans see no ads at all. For consumer brands this still leaves an enormous audience. For B2B it is the structural question — the senior buyers most worth reaching are the likeliest to be paying subscribers, outside the inventory entirely. That question deserves its own honest treatment, which it gets in our B2B read on ChatGPT ads.

The political question, answered plainly

Advocacy and public-affairs teams keep asking whether ChatGPT placement can carry an issue campaign. The honest planning answer is no, not today. Political advertising sits in the restricted bucket, the platform is cautious with sensitive categories while it is young, and no serious campaign should budget around inventory it probably cannot buy.

That is a smaller problem than it sounds, because the audiences an issue campaign needs are reachable through channels that have spent years building formal political-advertising lanes: programmatic display and video, connected TV, audio, and geofenced delivery around the places and moments where a policy audience actually is. Those lanes come with the verification, review and disclosure processes this work legally requires — mature compliance, rather than an absent policy. That is the mix we run as geofencing advertising for advocacy, and it is the practical answer for as long as assistant platforms keep political advertising gated.

Meanwhile the unpaid half of ChatGPT remains fully open to advocacy: being the source the assistant cites when someone asks about the issue. No ad policy gates citation. That work is AI search visibility, and for an issue campaign it is usually the higher-value lane anyway.

What to do with this

Three moves, in order. Check both lists before planning — business-country eligibility and user-country serving, against the platform’s current availability, because this guide’s snapshot will age. Declare the category honestly and expect manual review if you are anywhere near a regulated claim. And have the alternative ready: if the brief is restricted, the budget belongs in the channels with real political and regulated lanes, paired with the organic visibility work that no ad policy can gate.

Eligibility is the first step of every flight we scope: we confirm the brief against the current rules before any budget moves, and say so plainly when the answer is not yet. That check costs nothing — start here.