OpenAI Pushes ChatGPT Ads Into Southeast Asia and Taiwan, Passing 60 Countries

OpenAI says ChatGPT Ads is beginning to roll out across Indonesia, Malaysia, the Philippines, Singapore, Thailand, Vietnam and Taiwan, seven markets that the company describes as an extension of earlier Asia Pacific launches in Australia, New Zealand, Japan, South Korea and India. With the addition, OpenAI says the ad product is now available in more than 60 countries. The announcement is heavy on reach and partner names and light on the operational details advertisers usually ask for first.

What OpenAI actually announced

The mechanics described in the post are straightforward. Advertisers can buy through OpenAI's own Ads Solutions team, through agency partners that the company lists as dentsu, Havas Media, Omnicom Media, Publicis Groupe and WPP, or through unnamed technology partners. There is also self-service access via an Ads Manager, which OpenAI says is open to "eligible businesses" — the eligibility criteria were not spelled out.

On the consumer side, OpenAI repeats the policy it says applies in existing markets: ads appear only for people on the Free and Go plans, while Plus, Pro and Enterprise subscriptions stay ad-free. That is the clearest structural statement in the announcement, and it matters for anyone modeling the business. It sets a hard ceiling on inventory tied to the free and low-cost user base, and it creates an explicit incentive for OpenAI to keep that base large rather than convert it aggressively to paid tiers.

Dave Dugan, whom OpenAI identifies as vice president and head of global ads solutions, framed the expansion as reaching businesses of every size, from global brands to startups running a first campaign. Shopee's Peggy Zhu, executive director of brand and growth marketing, said the e-commerce company is extending a collaboration that began with the ChatGPT Ads launch in Brazil into Southeast Asia and Taiwan. No commercial terms, spend commitments or performance figures for that relationship were disclosed.

The revenue number, and what it does not tell you

OpenAI points back to an end-of-August disclosure that ChatGPT Ads had reached a $1 billion annualized revenue run rate in fewer than 200 days after launch, with tens of thousands of advertisers now running campaigns. That is a fast ramp by any standard, but annualized run rate is a projection built on a recent period, not a trailing revenue figure. It does not indicate whether spend is concentrated among a handful of large buyers, how much of it is experimental budget, or what renewal behavior looks like after a first campaign. OpenAI did not break out revenue by region, advertiser size or vertical, and did not say how the "tens of thousands" of advertisers are distributed across the 60-plus markets.

For product and growth teams evaluating the channel, the more useful missing numbers are the ones that never appear: no ad formats are described, no pricing or auction model is explained, no click-through, conversion or incrementality data is shared, and no third-party measurement arrangements are named. OpenAI says it will build "new formats, optimization tools, and measurement solutions" over the coming months, which is a reasonable roadmap statement and also an acknowledgment that the current toolkit is early.

The trust question sits at the center

OpenAI devotes a section of the announcement to its ads principles. The company says conversations are kept private from advertisers, that it never sells customer data, that ads are clearly labeled and kept separate from ChatGPT's answers, that advertising does not influence the answers the model gives, and that users have controls over ad personalization.

Those commitments address the obvious failure mode. A conversational assistant is a more intimate surface than a search results page: people describe budgets, constraints, health situations and business problems in detail. If users come to believe that a recommendation was purchased rather than reasoned, the utility of the assistant degrades quickly, and that erosion is hard to reverse.

The principles are stated as policy, not as an auditable mechanism. The announcement does not describe how separation between ads and answers is enforced technically, how personalization signals are derived if conversations are private from advertisers, what retention periods apply, or whether any external party reviews compliance. Those are the questions regulators in several of the new markets are likely to ask. Singapore, Indonesia, Vietnam, Thailand, Malaysia and the Philippines each have their own data protection regimes and, in some cases, localization requirements; the announcement does not mention regulatory review, local data handling or advertising-standards compliance in any of the seven markets.

Why Southeast Asia, and why now

The commercial logic is not stated in the announcement, but the shape of it is visible. These are markets with large mobile-first populations where free tiers dominate and subscription conversion is structurally harder than in North America or Western Europe. An ad-supported model is the mechanism OpenAI is using to fund access in exactly the places where paid plans are least likely to carry the load — a point the company makes when it says ads help support free and low-cost access to ChatGPT.

That also explains the Shopee reference. Commerce discovery is the use case where conversational advertising has the most obvious path to measurable outcomes, and Southeast Asia's e-commerce market is platform-concentrated in a way that makes a small number of partnerships disproportionately valuable.

The agency roster is the other tell. Listing the major holding companies signals that OpenAI is courting managed budgets rather than relying on long-tail self-service volume. That typically buys scale faster, but it also means early spend is easier to source than it is to retain, since agency test budgets move on when measurement does not hold up.

One more piece of context worth noting: the announcement arrives a day after OpenAI published posts on GPT-6 Sol and Luna and on prompt caching for GPT-6, and roughly a week after a piece titled "Reimagining advertising with AI." The ads business is being built alongside a fast-moving model cadence, and the announcement does not say how ad delivery interacts with newer models or how it is handled across the different plan tiers technically.

What to watch

• Whether OpenAI publishes ad formats, auction mechanics and pricing, or keeps buying concentrated through its own sales team and agency partners.
• Whether independent measurement and verification partners are named, since advertiser retention past initial tests usually depends on third-party validation rather than platform-reported metrics.
• How regulators in the seven new markets treat conversational ads, particularly on disclosure standards, personalization consent and data localization — none of which the announcement addresses.
• Whether the $1 billion run rate converts into disclosed trailing revenue, and whether OpenAI breaks out regional performance rather than country counts.
• Any signal that ad load or placement affects answer quality or user trust on the Free and Go tiers, which is the risk OpenAI's stated principles are designed to contain.