Singapore’s Gemini per-capita lead — and what it may mean for enterprise AI

Singapore is No. 1 globally for Gemini adoption per capita, according to Google’s internal data in its 2026 Gemini Southeast Asia report. It is a strong consumer-demand signal, but it does not yet tell us how Singapore firms use AI.

A follow-on question is how Singapore companies have redesigned work, put safeguards around staff use and embedded AI in core processes. A future enterprise companion report could make that second picture much clearer.

What happened

For the first quarter of 2026, Google’s first-party internal data ranks Singapore first globally per capita and first in Southeast Asia for daily engagement, at about 10 prompts per user per day. Users more than doubled in late 2025. People aged 25 to 34 made up about 40% of the local user base.

Singapore’s device mix also differs from the regional pattern. Google reports that nearly three in four Gemini prompts across Southeast Asia come from mobile devices; in Singapore, 58% came from computers, 40% from mobile devices and 2% from other devices. That may suggest Gemini is appearing in more desktop-based routines, but it does not tell us what happens inside firms.

These are first-party Gemini usage figures. They do not measure how firms govern staff use, choose models for a workflow, or integrate AI into core business processes.

What it means

Singapore’s Ministry of Manpower provides the more useful counterweight. In its 30 April 2026 release, MOM reported that 71.5% of firms had not adopted AI, while 28.5% had started. It recorded 7.4% planning adoption, 6.0% piloting and only 3.8% with AI integrated into core business processes.

MOM is measuring something different from Google: where firms sit in the work of adopting AI. Together, the figures suggest a market in which personal familiarity may be moving faster than organisation-wide integration.

Supporting context from AWS-commissioned research of SMEs in financial services, healthcare and manufacturing points in the same direction, but it is not a nationwide scorecard. Only 16% to 29% of AI-using SMEs across those sectors said they had reached advanced use, while fewer than three in ten said a clearly defined person oversaw AI accuracy.

An enterprise companion report would need to look beyond prompts and active users. It would need to show whether firms have usable rules for non-IT staff, select models for a defined workflow rather than treating a popular assistant as a default answer, and redesign processes with clear accountability and human review. The Straits Times reported NTUC deputy secretary-general Desmond Tan making a related point at the July launch of Singapore firm Synagie’s AI platform: firms should start with an operational problem, then decide whether AI is the right solution, alongside job redesign and skills investment.

Singapore may already have one of Southeast Asia’s most AI-familiar populations. Its next advantage will depend on whether companies turn that familiarity into better-designed work — with clear workflows, controls and accountable owners — rather than simply more prompts.

What to watch

• Firm movement: Whether MOM’s next adoption data shows more firms moving from planning and pilots into core-process integration.

• Operating evidence: Specific cases in which Singapore firms explain the workflow, controls and accountable owner behind an AI deployment, rather than only announcing a tool selection or adoption.

• Procurement discipline: Whether firms begin describing how they choose models by task, data sensitivity and oversight needs, especially as employees bring consumer expectations into work.