Singapore, AI and the visitor economy

Every domain listing claims perfect market timing. This one comes with sources, and with the arguments against it left in. Here is what is actually documented about AI and tourism in Singapore — and what it does not prove.

We built a Singapore AI concierge and we are selling the domain it runs on. That is an obvious incentive to tell you the market is inevitable. So instead of adjectives, below are the figures we could verify against primary sources, each one linked at the bottom so you can check it rather than take our word for it. Where a number is weaker than it looks, we say so.

The visitor base

Singapore recorded 16.9 million international visitor arrivals in 2025, up 2.3% on the year. The five largest source markets were China (3.1M), Indonesia (2.4M), Malaysia (1.3M), Australia (1.3M) and India (1.2M). Tourism receipts reached a record S$23.9 billion in the first three quarters of 2025 alone, up 6.5% year on year. The Singapore Tourism Board’s outlook for 2026 is 17–18 million arrivals and S$31.0–32.5 billion in receipts.

The honest footnote: 16.9 million came in slightly below STB’s own forecast range of 17.0–18.5 million for the year. Growth is real but modest, and it is not the vertical curve a pitch deck would prefer.

What matters more than the total, for a concierge specifically, is its composition. The top five markets speak Mandarin, Bahasa Indonesia, Malay, English and a range of Indian languages. A conversational product that only works in English is serving a fraction of the people arriving. This is why the concierge on this domain answers in whatever language the visitor writes in — and why that is a structural feature rather than a nice extra.

The institutional signal

In July 2025 the Singapore Tourism Board signed a memorandum of understanding with OpenAI — described as the first of its kind between OpenAI and a national tourism authority in Asia. Its scope covers personalised recommendations, multilingual assistance, immersive storytelling, and data analysis for demand forecasting and resource planning by operators. STB has also flagged AI concierges, robotics and autonomous tour vehicles as areas to explore under its Tourism 2040 roadmap.

A parallel signal from the private side: in April 2025 Singapore Airlines announced a partnership with OpenAI, reported as the first between OpenAI and a major global airline, to rebuild its virtual assistant on next-generation multimodal AI for both customers and staff. That assistant, Kris, has existed since December 2017 — it began as a modest pre-flight chatbot handling baggage and check-in questions. The distance between the 2017 version and what is being rebuilt now is a fair proxy for how far expectations have moved in this category.

The national AI backdrop

Singapore’s National AI Strategy 2.0 launched in December 2023 under the banner “AI for the Public Good, for Singapore and the World,” setting out 3 systems, 10 enablers and 15 actions. Among its commitments: tripling the country’s AI workforce from roughly 4,500 to 15,000 practitioners, and S$120 million for an “AI for Science” programme. A National AI Council chaired by Prime Minister Lawrence Wong was established in February 2026.

The most relevant piece for anyone building conversational products for this region is the National Multimodal LLM Programme — S$70 million from the National Research Foundation, run by IMDA with AI Singapore and A*STAR. It produced SEA-LION, an open-source model family covering 11 Southeast Asian languages, trained with roughly 13% Southeast Asian language data compared with about 0.5% in Llama-2.

That gap is the actual technical argument for regional AI products. Frontier models are excellent at English and thin on Bahasa, Tamil, Thai and Vietnamese, because the training data was thin. A destination concierge for Southeast Asia lives precisely in that gap.

The surrounding ecosystem is substantial: roughly 548 AI companies in Singapore including 9 unicorns, with about 143 funded and some US$1.9 billion raised collectively.

The .ai domain market

.ai registrations crossed one million at the turn of 2026, up from about 107,000 in 2020 — roughly a twentyfold increase in five and a half years. New registrations ran at about 1,318 a day through 2025 and about 2,008 a day in January 2026. (There is no authoritative count current to today that we could find, so we are not going to invent one.)

On the sales side, the clean data point is Bot.ai, which sold for US$1,200,000 via Sedo in February 2026 — the first publicly reported seven-figure .ai sale, roughly 60% above the previous .ai record, confirmed by the industry’s standard sales tracker. Beyond that one, a commercial tracker’s 2026 rankings place five .ai names in its top twenty sales for the year, in a range from US$275,000 to US$1.2 million; we could not independently corroborate the four besides Bot.ai, so treat those as one source’s reporting rather than established fact.

Two counterweights worth stating. Premium .com still dominates the top-sales tables — .ai is a strong second, not the leader. And headline sales are, by construction, the visible tail of a distribution whose middle nobody publishes. A record sale tells you the ceiling exists; it tells you nothing about any particular name.

One genuinely odd fact, since it comes up: .ai is the country-code domain for Anguilla, and its registration revenue now funds a material share of that island’s national budget — airport works, roads, tax relief, health services. An AI boom paying for Caribbean infrastructure is not a business argument, but it is true, and it is a reminder that this extension has an owner and a history rather than being a marketing invention.

What this does not prove

Attention is not demand. Every figure above describes governments, airlines and registries moving toward AI — none of them describes a visitor paying for an AI concierge, because that behaviour is not yet established at scale. The honest summary is that the enabling conditions are unusually good in Singapore specifically: high visitor volume, multilingual demand, strong digital infrastructure, an explicit national strategy, a tourism authority publicly exploring exactly this, and clear data-protection rules to build inside.

Enabling conditions are not a business. They are the reason a working proof of concept on the right name is worth someone’s attention now rather than in three years, when the good names in this convention are gone and the question is which of your competitors took them.

You can test the proof of concept yourself in about thirty seconds: ask it something. It will answer, in your language, and it will tell you when it doesn’t know — which, as we wrote elsewhere, was the difficult part.

Sources

• Singapore Tourism Board — STB and OpenAI sign first-of-its-kind MOU (July 2025)
• Singapore Tourism Board — Record tourism receipts, January–September 2025
• TTG Asia — 2025 visitor arrivals and 2026 forecast (February 2026)
• Singapore Airlines — SIA–OpenAI collaboration (April 2025) and the launch of Kris (December 2017)
• Smart Nation Singapore — National AI Strategy 2.0
• AI Singapore — SEA-LION and the National Multimodal LLM Programme
• Singapore AI company census
• DNJournal — Bot.ai sale report (February 2026) · Instant Domain Search top sales 2026
• Hogan Lovells — .ai reaches one million registrations