Southeast Asia has secured more than USD 55 billion in AI investment commitments, yet the ten-member bloc still has no binding regional framework covering data sovereignty, technology transfer, or investment terms. The ASEAN Guide on AI Governance and Ethics, endorsed by ASEAN digital ministers in February 2024, remains entirely voluntary. Meanwhile, Singapore alone attracted around USD 2.3 billion in AI-related deals between 2019 and mid-2024 — a concentration that is accelerating regional divergence rather than collective strength.
The structural risk is not that Southeast Asia lacks AI investment. It is that without coordinated strategy, individual countries are negotiating separately with the world’s most powerful technology firms and losing leverage with every deal they sign alone.
Southeast Asia is on course to build a USD 600 billion digital economy by 2030, yet the infrastructure decisions being made right now — on compute access, data governance, and model deployment — are largely being made by foreign firms, not regional governments. Khailee Ng, a partner at venture capital firm 500 Global, put the danger plainly on the Keris & Silicon podcast: countries across the bloc are “competing among themselves, doing their own thing, and not bundling much negotiating power to really exercise leverage in very important deals.”
The consequence is structural, not rhetorical. Malaysia, Indonesia, and Thailand have each secured bilateral AI investment pledges — USD 2.2 billion, USD 1.7 billion, and USD 2.7 billion respectively from major technology firms — but negotiated independently, each government sitting across the table from the same hyperscalers without a common position. Andrew Ng, founder of Deeplearning.ai and AI Fund, made the same point at the 2025 AI for Good Global Summit: without local investment in talent and compute, emerging economies will capture only “a small fraction” of AI’s value chain, becoming downstream consumers of systems designed and owned elsewhere.
That is the buried story beneath the investment headline numbers.
RELATED
Fragmented bets, concentrated returns
The Google, Temasek, and Bain & Company e-Conomy SEA 2024 report projects Southeast Asia’s digital economy reaching USD 600 billion in gross merchandise value by 2030. The AI-specific investment picture, however, is far narrower: Singapore captured the overwhelming share of AI-related capital flows across the region between 2019 and mid-2024, with approximately USD 2.3 billion in deals — while Indonesia and Vietnam, home to the bloc’s two largest populations, are scaling startup activity on imported models and rented cloud compute.
Singapore’s lead is deliberate. Its National AI Strategy 2.0, released on December 4, 2023, targets more than tripling the country’s AI workforce to 15,000 specialists and practitioners and deploying AI across at least 15 high-impact national projects within three to five years. Ng Chee Khern, Chair of the Strategy 2.0 Steering Committee, framed compute, data, and talent explicitly as “strategic national assets” in March 2024 — language that contrasts sharply with how most of Singapore’s neighbours have approached the same question.
The ASEAN-level response has not matched that ambition. The ASEAN Guide on AI Governance and Ethics, while a meaningful starting point, carries no binding obligations and no enforcement mechanisms. The gap between that voluntary framework and the EU’s Artificial Intelligence Act — formally adopted in March 2024, phased in between 2025 and 2027, with mandatory conformity assessments and prohibitions on high-risk uses — is not merely regulatory. It is a bargaining power gap.
| Country | Reported AI investment pledge | Framework context |
|---|---|---|
| Thailand | USD 2.7 billion | No binding ASEAN standard; bilateral terms undisclosed |
| Malaysia | USD 2.2 billion (Microsoft) | Negotiated separately; no regional data-sovereignty clause |
| Indonesia | USD 1.7 billion | Bilateral deal; no cross-border technology transfer terms |
China’s approach offers an instructive contrast. Chinese state media reported in February 2026 that the country had 602 million generative AI users as of December 2025 — up 141.7% from the end of 2023 — with 254,000 new AI-focused SMEs registered in the first quarter of 2025 alone. That adoption rate was driven in part by Chinese technology firms embedding large language model capabilities directly into consumer products, often at no cost, to accelerate mass uptake. Southeast Asian governments have no equivalent industrial-policy mechanism to push comparable adoption at scale.
Why the infrastructure layer is the decisive bet
The Asia Development Bank Institute estimated in a 2025 report that generative AI could add between 0.5 and 1.0 percentage points to annual GDP growth in more digitally advanced ASEAN members by 2030 — but only if accompanied by complementary investment in skills and infrastructure. That conditional clause is doing significant work. The countries best positioned to capture those gains are precisely those already investing in sovereign compute capacity, AI talent pipelines, and governance frameworks that give them credibility in negotiations with global technology firms.
The competitive landscape makes the stakes concrete. The global AI stack is dominated by US hyperscalers — OpenAI-Microsoft, Google, Amazon, and Meta — with Baidu, Alibaba, Tencent, and ByteDance as major rivals. Singapore hosts regional AI hubs for many of these firms and is cultivating applied AI through Sea Group and Grab. Indonesia and Vietnam have dynamic startup scenes but rely heavily on imported models and rented cloud compute. That reliance is not neutral: it means pricing, access conditions, and technical roadmaps are set elsewhere, and local governments have limited recourse when terms change. The infrastructure layer is where sovereignty is actually decided — not in strategy documents.
Watch for the outcomes of the ASEAN Digital Ministers’ Meeting expected in late 2026, when ministers are due to consider implementing the 2024 AI Governance Guide. If they agree on concrete mechanisms — shared testing facilities, cross-border data arrangements — it signals genuine coordination is emerging. If they do not, bilateral experimentation continues and collective bargaining power with hyperscalers narrows further.
Beyond the headline
The Bigger Picture
Southeast Asia’s uneven AI readiness is part of a broader pattern where the region excels in consumer-facing digital services but underinvests in foundational infrastructure and standards. That imbalance risks locking local economies into low-margin implementation roles while value and decision-making migrate to jurisdictions that control chips, clouds, and rules.
The Power Behind It
Although regional leaders talk about digital sovereignty, practical control over AI deployment still sits largely with a handful of global cloud and semiconductor firms. Their pricing, access conditions, and technical roadmaps quietly shape what is realistically possible for Southeast Asian governments, often more than any local policy statement or strategy document.
The Reach
For Western enterprise software vendors, Southeast Asia’s fragmented AI governance means navigating a patchwork of requirements rather than a single ASEAN standard. That fragmentation raises compliance costs but also creates opportunities: firms that can align their offerings with Singapore’s relatively advanced frameworks may set de facto norms other countries eventually follow.
What Southeast Asia’s AI fragmentation means for your decisions now
With no binding ASEAN AI framework in place and the next ministerial review not expected until late 2026, the governance gap is a live operational variable — not a future risk.
-
Technology investors and fund managers
The concentration of AI capital in Singapore — approximately USD 2.3 billion in deals between 2019 and mid-2024 against a regional total of over USD 55 billion in commitments — signals where infrastructure-layer returns are most likely to compound. Indonesia and Vietnam offer earlier-stage exposure to applied AI and talent growth, but compute dependency on foreign providers is a structural risk to margin. Track Singapore’s budget announcements on data centre and GPU funding as a leading indicator of whether regional anchor infrastructure holds.
-
Enterprise technology and software vendors
Compliance planning across Southeast Asia currently requires country-by-country assessment: Singapore’s Personal Data Protection Commission advisory guidelines and Model AI Governance Framework are the most developed in the bloc, but they are voluntary. The EU AI Act’s phased obligations between 2025 and 2027 provide a higher baseline for any vendor already operating in Europe. Firms that build to Singapore’s standard first are best placed to adapt as other ASEAN members develop harder rules — and to influence what those rules look like.
-
Policy professionals and government advisers
The ASEAN Guide on AI Governance and Ethics, endorsed in February 2024, is the current regional reference point — available directly from ASEAN’s official publication portal. The gap between that voluntary framework and binding instruments like the EU AI Act is the central policy design question for the bloc over the next two years. The late-2026 ASEAN Digital Ministers’ Meeting is the next concrete decision point; monitoring its agenda and outcomes will indicate whether collective bargaining with hyperscalers becomes viable or whether bilateral deal-making remains the default.
