OpenAI announced 26 additional ChatGPT Ads countries on October 21, 2026, bringing its stated total to 89. For advertisers, the useful next step is to reassess a shortlist of markets, not spread an existing budget across every newly available country. This article was reviewed on October 26; the expansion announcement is dated October 21.
What the announcement establishes
The expansion includes Cambodia and Colombia, among other markets. OpenAI describes access through technology partners or self-service Ads Manager. Its plans for more formats, objectives and measurement are future direction, not confirmation that every account has those capabilities.
Read the October 21 announcement
Check the buyer before choosing the market
The current Ads Manager availability table lists Cambodia, Colombia and Vietnam as available. Self-service eligibility depends on where the advertising and billed legal entity is based; creating an account alone does not establish access. Check the live table and your account before making a launch commitment.
Check current Ads Manager availability
A market-selection exercise, not a rollout guarantee
The following is GPTAd editorial guidance. Build a one-page decision record for each serious candidate rather than treating the country total as a media plan.
1. Define the commercial reason. Name the product, customer problem and evidence of demand in that market. Existing qualified inquiries or repeat customers are a stronger starting point than availability alone.
2. Write down the disqualifiers. For example: no delivery coverage, an unsupported checkout currency, no one to answer inquiries, or insufficient margin after fulfillment. A candidate that fails a critical operational check stays out of the first test.
3. Specify the evidence a test must produce. Choose a business outcome such as qualified inquiries or completed purchases. Decide who reconciles it with sales records, and record the approved spending ceiling and review date. These are your internal decisions, not platform-mandated thresholds.
4. Assign a decision owner. The person approving expansion should be able to stop the experiment, investigate poor lead quality and distinguish an untested market from an unsuccessful one.
Example: more reach does not require more simultaneous tests
Consider a hypothetical software company choosing between two newly accessible markets. One already produces relevant sales inquiries and has a support owner; the other has only general website traffic. A reasonable first experiment could prioritize the former while interviewing potential customers in the latter. That recommendation follows the company’s evidence, not a claim that either country has cheaper clicks or better conversion rates.
Keep each test interpretable. Changing market, language, offer and success criteria together makes it harder to explain the result. Record the assumptions before launch and review the customer experience alongside the advertising report.
Use the first-campaign learning plan
Vietnam was already covered by the earlier regional expansion. Teams operating from Vietnam can use the existing eligibility guide as background; the October 21 announcement should not be presented as Vietnam’s first launch.
Read the Vietnam eligibility guide
GPTAd is an independent service and editorial publisher. This planning framework is not an OpenAI endorsement, account approval or forecast of campaign performance.
