e-Conomy SEA 2026: What It Says About AI Use in Indonesia

What does the e-Conomy SEA 2026 report say about AI use in Indonesia, and what does it mean for businesses? The 11th edition of the annual study by Google, Temasek and Bain & Company, titled Building the Intelligent Economy and released on 21 October 2026, finds that generative AI has spread across Southeast Asia faster than mobile internet did, but that most people use it lightly and most use still comes from consumers rather than organisations. For Indonesia, local coverage of the report puts consumer use at about 80% of the country's AI activity. That gap between personal use and use inside companies is the most practical finding for Indonesian firms planning their AI work.

The headline figures

According to the press release, Southeast Asia's digital economy is projected to reach US$368 billion in gross merchandise value in 2026, up 17% on the previous year, and to pass US$1 trillion by 2035. Bisnis.com, reporting on 27 October, gives the 2035 figure as US$1.13 trillion, a compound annual growth rate of 13%. E-commerce remains the largest sector at US$226 billion, up 18%. Digital financial services are put at US$1,650 billion in gross transaction value, up 16%, with digital lending balances up 22% to US$116 billion.

On AI, the report says 243 million people, or 58% of the working population, have adopted generative AI in under three years, about four times faster than mobile internet spread. This is the first edition to measure AI use in tokens, the units of text that AI models process, which the authors treat as the AI equivalent of gross merchandise value. Measured that way, use per person is roughly 10% of US levels, and only about 1 to 2% of people in the six largest markets pay for generative AI, against 3% of US households. Florian Hoppe, a partner at Bain & Company, said the region's AI story is shifting from how many people use it to how deeply they use it.

What the report says about Indonesia

The regional press release and Google's summary contain few Indonesia-specific figures, and we did not find a 2026 digital economy value for Indonesia in the materials we read. Bain's summary says Indonesia and the Philippines have the most room for future growth in AI users. Bisnis.com reported several country findings from the report:

• About 54% of Indonesia's working population now uses generative AI, slightly below the regional figure of 58%.
• Consumers account for about 80% of Indonesia's AI token use, compared with about 60% across the region.
• Indonesian users stand out for creative use: 32% of their queries on Google's Gemini are used to generate images, the highest share in Southeast Asia, at about 9 million images a day.
• Indonesia has 17 digital banks operating since 2021, the largest such group in the region, and some, including SeaBank and Bank Jago, are profitable.
• Planned AI infrastructure in Batam includes a US$5 billion high-density AI data centre announced by Equator Gate System, and a 360MW AI campus for about 170,000 GPUs announced by Firmus, which is backed by NVIDIA.

Youngster.id, reporting on the release, said the report describes AI use in Indonesian finance as focused on credit scoring to reach people without bank accounts and on real-time detection of online fraud.

Where organisations fit in

The report defines enterprise workloads as AI use that organisations buy, deploy or manage. It says these are strongest in the more mature markets, making up 45 to 60% of use in Singapore and Malaysia, and that enterprise use has significant room to grow across the region. Set against Bisnis.com's figure of about 80% consumer use in Indonesia, this suggests a smaller share of Indonesia's AI activity happens inside managed company systems, though the figures come from different parts of the report and may not be measured in exactly the same way.

Investment is moving towards AI quickly. Companies with core AI products took 80% of private funding deal value for digital applications in the last six months, up from 30% in the second half of 2024. AI compute and digital infrastructure drew US$20 billion of private capital over the past 12 months, led by the US$11 billion acquisition of ST Telemedia Global Data Centres. The report says regional data centre capacity needs to grow about threefold over five years to meet demand.

The conditions attached to the forecasts

The report projects that regional AI use could grow up to 60-fold by 2030, but ties that to four conditions: wider use of multi-step agentic workflows, in which AI carries out a sequence of tasks rather than answering one question; more AI infrastructure; more efficient models and energy use; and falling inference costs alongside demonstrable returns on investment. Tech Wire Asia noted that the token measure is new, so there is no earlier baseline to compare against, and that Google, a co-author, sells AI models and cloud capacity. These are projections, not measured outcomes.

The survey findings on trust are also conditional. According to Google's summary, 72% of users already rely on everyday automations such as price alerts, 75% would let an AI agent find the best price and complete a purchase for them, and 62% trust AI more than they did a year ago. But respondents named conditions: privacy safeguards (44%), clarity on what the AI can and cannot do (36%) and the ability to reverse an agent's actions (30%).

What businesses can take from it

This section is our analysis. If most AI use in Indonesia happens on personal accounts, staff are already using AI at work in ways the company may not see or govern. The step the report points to, from casual use to managed workflows, is mainly a design and data task rather than a question of buying a model.

• Find out where staff already use AI tools for work, what company data goes into them, and which of those tasks would justify a managed tool with access controls and logs.
• Pick one or two multi-step processes, such as handling customer documents or answering internal questions from policies, and measure the result against an agreed test set before widening use.
• If customers will deal with AI agents, build in the conditions the survey names: clear limits on what the agent can do, privacy safeguards, a record of its actions and a way to reverse them.
• Track cost per task from the start, since the report links future growth to falling inference costs and demonstrable returns.

What to watch next

Watch the signing of the ASEAN Digital Economy Framework Agreement, which the report expects in November 2026 and lists among the conditions for regional growth. Watch whether the Batam data centre projects proceed on the scale announced, since local capacity affects where Indonesian companies can run AI workloads. And watch next year's edition for the enterprise share of AI use in Indonesia: in our view, that number, more than the count of users, will show whether AI is changing how Indonesian organisations work.