Local languages and mobile use drive Gemini’s growth across Southeast Asia

Google’s Gemini app more than doubled its monthly active user base across Southeast Asia between April 2025 and March 2026, with younger users and local-language interactions accounting for much of the activity recorded in its first regional usage report.

Nearly 70% of prompts across Indonesia, Malaysia, the Philippines, Singapore, Thailand and Vietnam were submitted in local languages. Almost three-quarters of requests came from mobile devices, while voice, photographs and video appeared in more than 40% of prompts.

Together, those patterns show how people are fitting the AI assistant into existing device and language habits. Many interactions involve quick tasks carried out by phone, including translating signs, reviewing documents and asking for help with everyday problems.

The Gemini Report: Southeast Asia 2026 is based on Google’s internal data, with most of its recent usage findings drawn from the first quarter of 2026. It covers the Gemini web and mobile apps but excludes use through other Google products such as Search and Workspace.

Local-language support broadens how people use Gemini

Local-language prompts accounted for 89% of Gemini requests in Vietnam, 87% in Thailand and 84% in Indonesia. The high proportions indicate that many users are approaching the assistant in the language they already use for work, study and daily communication.

AI Singapore’s Southeast Asian Holistic Evaluation of Language Models, known as SEA-HELM, also ranked Gemini as the best-performing large language model overall for Southeast Asian languages. The evaluation examines how models handle the region’s linguistic characteristics and cultural contexts.

That capability affects more than the wording of a prompt. Users who can speak or write naturally in their preferred language can also make greater use of voice, images and video without translating their request into English first.

Voice-only interactions accounted for more than 10% of prompts across the region. Other users uploaded photographs or video when they needed help understanding a document, identifying something in front of them or working through a repair.

Younger users were the most engaged group. People under 25 submitted more requests, held longer conversations and wrote more detailed prompts than older users, according to Google’s data.

Their activity extended across creative and information-based tasks. Around 40% of classified prompts involved generating new material, including writing, images, music and video. Other common uses included summarising documents, organising information, researching topics and seeking recommendations.

Device choice shapes the type of work

Mobile devices dominated Gemini use in five of the six markets covered by the report. They accounted for 82% of prompts in Indonesia, 74% in Malaysia and Thailand, 71% in the Philippines and 66% in Vietnam.

These sessions tended to involve shorter conversations, image generation, casual questions and quick requests for advice. Computer users were more likely to work through longer prompts and tasks that required several exchanges.

Singapore differed from the regional pattern. Computers generated 58% of requests, compared with 40% from mobile devices, making it the only market where desktop use exceeded mobile activity.

The country also recorded the highest per-capita Gemini adoption globally and the highest daily engagement among the six Southeast Asian markets, based on Google’s internal figures. People aged 25 to 34 made up 40% of its daily user base.

Usage in Singapore followed the working week. Requests involving explanations, research and code debugging were more common on weekdays, when computers accounted for 60% of prompts. At weekends, casual conversations and lifestyle requests rose to almost one-third of activity, with travel and personal finance appearing more frequently than in the other markets studied.

Google adds task execution through Gemini Spark

The report’s usage patterns provide the backdrop for Google’s expansion of Gemini Spark, a cloud-based AI agent that can complete tasks through Gmail, Docs and Slides.

Spark can continue working while a user’s computer is closed or phone is locked. It is currently available in English to Google AI Ultra subscribers, with support for local Southeast Asian languages beginning this week.

The rollout builds on the types of work already appearing in Gemini conversations, including research, document handling and content creation. Spark extends those activities across Google’s other services, where the assistant can act on information rather than respond within a single chat.