Google says Gemini app users in Southeast Asia have more than doubled in a year

Google has published its first Gemini Southeast Asia Report, arguing that the region has become one of the clearest tests of whether consumer AI assistants can move beyond English-first chat and into everyday mobile use.

In a July 14 post, Google said active Gemini app users in Southeast Asia have more than doubled over the past year and that people in the region are adopting Gemini faster than any other Google app. The company focused its analysis on six large user bases: Indonesia, Malaysia, the Philippines, Singapore, Thailand and Vietnam.

Why the regional data matters

The most important detail is language. Google says nearly 70% of prompts in the region are submitted in native languages rather than English, with Vietnam at 89%, Thailand at 87% and Indonesia at 84%. That helps explain why voice, images and video are central to the usage pattern. Google says almost three in four Gemini requests in Southeast Asia come from mobile devices, while more than 40% of prompts include voice commands, photos or video uploads. Voice-only use accounts for 10% of prompts.

Those figures point to a different AI growth story than the desktop productivity narrative common in the United States and Europe. In Southeast Asia, Google is presenting Gemini as a mobile-first assistant for multilingual users who may be more likely to ask, show or speak than type a carefully formatted prompt.

Spark expands in local languages

Google is also using the report to push Gemini Spark deeper into the region. The company says Spark will roll out in local languages to Gemini Advanced Ultra subscribers this week, following an initial English-language launch. Spark is described as a more proactive agent that can work across Google Workspace tools such as Gmail, Docs and Slides, including background task handling when a device is locked or closed.

Google also highlighted creative usage. It said roughly 40% of Gemini queries in the region ask the assistant to generate new output, including images, music, video and documents. The company said users in Southeast Asia have generated 5 billion images with its Nano Banana image model over the past year and almost 1 million songs since Lyria 3 was introduced in the region.

The claims come from Google's own product data and should be read as company-reported metrics. Still, the report is useful because it shows where the next phase of AI assistant competition may be fought: not only in model benchmarks, but in local-language fluency, mobile multimodal workflows and agent features that fit regional habits.

Sources

• blog.google

Cover photo by Ramius Aquiler on Pexels, used under the Pexels License.