ORAC

----

Singapore a high-value market for ChatGPT ads, but Indonesia, Vietnam offer scale: analysts

LongbridgeAII'm LongbridgeAI, I can summarize articles.

Analysts note Singapore offers high-value users for ChatGPT ads, while Indonesia and Vietnam provide greater scale due to larger populations. OpenAI launched ads in Singapore and other SEA markets to monetize its free user base, aiming for $2.5B ad revenue this year. Experts warn that trust issues may arise as ads appear in conversational AI, potentially impacting user perception. This expansion intensifies competition with traditional search engines like Google.

[SINGAPORE] Singapore may offer OpenAI some of South-east Asia’s most valuable users for its nascent advertising business, but larger markets such as Indonesia and Vietnam could ultimately offer greater scale as the ChatGPT maker looks to monetise its large free-user base.

OpenAI on Sep 24 began rolling out ChatGPT ads in Singapore, five other South-east Asian markets and Taiwan, extending a new revenue stream beyond subscriptions.

Ads are shown to ChatGPT users on its free tier and its lowest-cost Go plan, which costs S$11 a month in Singapore.

Li Jianggan, CEO of market insights firm Momentum Works, noted that Singapore is one of the most valuable ad markets in South-east Asia because of its higher purchasing power and advertiser value per user.

This higher purchasing power, however, means that ChatGPT users in the Republic are also more willing to pay for subscriptions, he said. This potentially shrinks the pool of users available to advertisers.

While OpenAI does not disclose granular market data, an OpenAI spokesperson told The Business Times that Singapore has one of the highest per-capita usage rates of ChatGPT.

Singapore has the second-highest artificial intelligence penetration rate globally among the working-age population at 64.3 per cent, according to Microsoft’s Global AI Diffusion Report released in September.

Indonesia and Vietnam have considerably lower penetration rates, at 14.8 per cent and 27.9 per cent, respectively, but their much larger populations give OpenAI a substantially bigger potential pool of users to monetise.

Li said that the gap between ChatGPT usage and consumers’ willingness to pay for an AI subscription is likely to be larger in these markets, increasing the importance of advertising as an alternative source of revenue.

“So Singapore may be a better advertising market on a per-user basis, (and) Indonesia and Vietnam could ultimately be more important in terms of scale,” he told BT.

More broadly, Li said that having ads in the AI large language model (LLM) makes sense for South-east Asia, as it enables OpenAI to monetise a much larger free-user base without requiring them to pay for subscriptions.

Dr Samer Elhajjar, senior lecturer at the National University of Singapore (NUS) Business School, added that Asean consumers tend to be highly adaptable when it comes to using new technologies, which could make them more receptive to advertising on AI platforms.

“In that sense, they may be more accepting of ads within AI platforms.”

Trust in the output

But analysts cautioned that acceptance of advertising does not necessarily translate into trust.

OpenAI has said that ads will be clearly labelled and differentiated from actual responses, and will not influence outputs.

Still, Dr Elhajjar said that the distinction may not be so clear-cut from a user’s perspective. “Even if an ad is clearly labelled, it still appears within a conversational environment that users trust.”

The introduction of ads might even damage users’ trust in the answers themselves, “because trust is driven as much by perception as reality”, he added.

Li similarly cautioned against assuming South-east Asian consumers to be accepting of ads, simply because super-apps and e-commerce platforms are already ad-heavy.

“The more important issue is whether the ad is useful in the context of the conversation,” he said. “ChatGPT is quite different from scrolling TikTok or browsing Shopee, because users may perceive the assistant as helping them make a decision.”

A new revenue stream

The expansion comes as AI companies face pressure to turn user growth into sustainable revenue, noted analysts, with OpenAI seeking a potential initial public offering in 2027.

“Every free user costs real money in computing power,” said Dr Zhang Jiehuang, a lecturer at the College of Computing and Data Science at Nanyang Technological University (NTU).

As such, advertising is the “obvious way” to pay for any assistant with a large free user base, he added.

With the addition of South-east Asia and Taiwan, ChatGPT now has advertisement services in more than 60 markets, bringing annual recurring revenue to more than US$1 billion since its launch in February this year.

OpenAI told investors that it expects to generate US$2.5 billion in advertising revenue this year, and projects to reach US$100 billion by 2030, according to media reports.

The emergence of advertising within AI chatbots could also intensify competition with traditional search engines – such as Google – for digital advertising dollars over time, said analysts.

A 2024 report from research and advisory firm Gartner forecast that traditional search volume would drop 25 per cent by 2026, as generative AI became an “alternative answer engine”.

Joe Escobedo, an adjunct lecturer at NUS Business School, said that such AI models will act as the most direct threat to Google’s business model in analysing search queries.

Concurring, Li said: “Over time, Google Search is probably the most directly comparable pool because ChatGPT increasingly captures explicit consumer intent.”

Nevertheless, Escobedo pointed out that search and related earnings for Google parent Alphabet continue to surpass those of ChatGPT, with Alphabet generating US$198.1 billion in 2024.

More ads in LLMs

Analysts remain split on the future of advertising within AI chatbots, as it will depend on how the individual companies choose to monetise their services.

Dr Elhajjar noted that “GenAI is expensive to operate, and investors will eventually expect sustainable revenue models”.

“Advertising is one of the most established ways of monetising digital platforms,” he added.

That said, he believes that not every AI company will copy the “traditional advertising playbook”, as some companies will integrate commerce, affiliate recommendations, premium placements or branded services instead.

NTU’s Dr Zhang sees a move towards “two business models: assistants paid for by advertisers and assistants paid for by users or employers”.

“That is fine, as long as people can tell which kind they are using and what implications come with it.”

Login to unlock5,578characters for free

Due to copyright restrictions, please log in to your Longbridge account to view this content.
Thank you for your understanding and support of licensed content.