View from space of illuminated cities at night

Why sovereign artificial intelligence is imperative in Southeast Asia

Authors

Contributors

5 minute read 17 Apr 2026

While Southeast Asia has taken meaningful steps in artificial intelligence (AI), a shared, operational sovereign AI model is needed.

In brief:

  • AI systems trained, updated or inferred on platforms beyond national oversight create a sovereignty gap.
  • Sovereign AI allows governments and enterprises to innovate with confidence while safeguarding trust, safety and resilience.
  • A coordinated ecosystem is also essential as no single party can deliver sovereignty end‑to‑end.

Across Southeast Asia (SEA), artificial intelligence (AI) is moving rapidly from pilot initiatives to the core of how governments, large enterprises and regulated industries operate. Sovereignty is not simply about where AI systems are hosted; it is about maintaining meaningful control over how AI systems behave, make decisions and evolve over time. As AI increasingly affects citizens, customers and mission‑critical operations, organizations need assurance that these systems operate according to domestic frameworks, not assumptions embedded in foreign platforms.

The ASEAN Guide on AI Governance and Ethics (2024) and the Expanded Guide for GenAI (2025) provide shared principles — accountability, assurance, testing and provenance — for cross-border interoperability and trust. Markets in the region have also taken meaningful steps to define responsible AI principles, update privacy laws and launch assurance initiatives.

For example, Singapore has operationalized AI assurance at scale through AI Verify and the Global AI Assurance Sandbox, allowing real-world testing of generative AI systems. Malaysia issued The National Guidelines on AI Governance & Ethics (2024) and established the National AI Office to coordinate implementation and upcoming legislation. The Philippines issued AI guidance (2024) clarifying application of the Data Privacy Act across training, testing and deployment. In 2025, Indonesia’s constitutional court clarified mandatory data protection officer requirements, tightening accountability for high-risk processing and AI‑enabled operations. Vietnam enacted the region’s first standalone AI Law (effective March 2026).

However, progress remains uneven across the region, with operational gaps consistently appearing in three areas.

How EY can Help

1. Where models run

High-risk AI workloads are still often executed on foreign-hosted platforms. Without explicit in‑country requirements, inference, update cycles and model behavior data can fall outside domestic oversight — a direct challenge to sovereignty.
 

2. How decisions are monitored

Requirements for runtime logging, activity records, reviewable decision trails and regulator‑ready evidence are evolving unevenly. Assurance remains fragmented, creating uncertainty for regulators and operators in high-stakes environments.
 

3. How AI stays aligned in production

Most frameworks explain how to assess AI risk, but they rarely show how to keep AI aligned once it is deployed. Organizations still lack clear guidance on continuous monitoring, testing and adjustment of models in live environments, especially when systems are distributed or use sensitive data.

 

These gaps limit the safe use of AI in sensitive domains, such as citizen services, finance, healthcare, energy and border operations.

Sovereign AI in Southeast Asia

This EY sovereign AI report highlights practical guidance for governments and enterprises to close AI sovereignty gaps without replacing existing infrastructure and core systems.

Request a copy

How EY can Help

Why sovereign AI is key 

Sovereign AI means that the models, decisions and behaviors influencing critical services and national outcomes remain governed, monitored and aligned within national boundaries. It also encompasses control over data, infrastructure, computing, guardrails, update cycles and the operating environment — ensuring sovereignty across the full lifecycle of how AI is developed, deployed and hosted. In addition, it provides a framework where data, models and decision‑making stay explainable, auditable and under domestic oversight, enabling responsible innovation at scale.

 

Sovereign AI addresses the gaps mentioned earlier in several critical areas.
 

1. Compliance cost

As hosting, logging and audit requirements tighten, non-sovereign architectures might face costly retrofits, migrations and operational disruptions. Sovereign AI helps reduce the cost of such compliance.
 

2. Data protection and intellectual property control

For enterprises, sovereign AI is relevant as it strengthens data protection and intellectual property control. Many enterprises rely on sensitive data that ranges from customer or client information to proprietary research. In financial services, for example, banks handle vast amounts of personal and transactional data that must comply with strict local data protection laws. Sovereign AI capabilities enable financial institutions to keep data within controlled environments, minimizing risks of breaches and ensuring compliance. This is particularly important as consumers demand transparency and robust privacy protection.
 

3. Operational resilience

As businesses integrate AI into supply chains, customer engagement and decision-making, dependence on external or foreign platforms can create vulnerabilities. Sovereign AI enhances operational resilience as it gives enterprises greater control over AI infrastructure and deployment environments, helping to reduce enterprise exposure to disruptions.
 

4. Competitive advantage

Sovereign AI also creates a competitive advantage. Enterprises that can tailor AI models to their own data, processes and industry needs gain unique insights and efficiencies that generic, one-size-fits-all AI solutions cannot provide. For instance, in healthcare, a medical research institution can use sovereign AI to analyze patient data and develop targeted treatment protocols. This helps drive healthcare innovation within clear governance frameworks.

Need for a coordinated ecosystem

Sovereign AI requires a coordinated ecosystem. No single party can deliver sovereignty end‑to‑end. The organizations and economies that can align infrastructure, computing, models and governance internally may carry a premium of control. They will be able to innovate safely at scale, withstand external shocks and shape the standards others follow. Achieving this level of coherence requires clarity, capability and partners who can help translate sovereign intent into day‑to‑day operational reality.

The EY DecisionOS platform provides this capability through a domestically hosted decision layer that orchestrates data flows, model execution, decision formation and runtime oversight across existing systems without major platform replacement. Beyond integration, DecisionOS enables sovereignty to be exercised in day‑to‑day operations. It establishes a shared decision fabric where infrastructure, computing, sovereign‑compatible models and enterprise applications operate as a coordinated, domestically governed system.

In doing so, DecisionOS becomes the operational backbone for sovereign‑grade AI, ensuring that every decision — automated or human‑in‑the‑loop — is fully traceable, governed and aligned to domestic policy requirements.

By embedding governance, auditability and control into the decision lifecycle itself, it enables organizations to deploy advanced AI capabilities with confidence, knowing that data residency, model behavior and system interactions remain fully under national jurisdiction. This allows institutions to scale AI‑driven transformation while maintaining the integrity, security and autonomy expected of a sovereign digital ecosystem.

Summary

Sovereign AI is essential as it enhances data protection, intellectual property control and operational resilience while helping to reduce compliance cost and drive innovation within clear governance frameworks. However, it also requires a coordinated ecosystem as no single party can deliver sovereignty end‑to‑end. The EY DecisionOS platform can address this need.

Related content

How organizations can capitalize on the rise of AI factories

Organizations must shift from capability-driven AI to factory-style intelligence to drive enterprise-level performance. Learn more.

Manik Bhandari + 1

How can responsible AI bridge the gap between investment and impact?

Explore the ways in which responsible AI converts investment into meaningful impact.

Joe Depa

How data analytics and AI in government can drive greater public value

Discover how data analytics and AI in government can drive greater public value and learn lessons from successful government organizations. Find out more.

Sheena Thompson + 1

    About this article

    Authors

    Contributors

    Related topics
    Consulting
    AI
    Transformation
    Technology
    Technology leader's agenda

    EY refers to the global organization, and may refer to one or more, of the member firms of Ernst & Young Global Limited, each of which is a separate legal entity. Ernst & Young Global Limited, a UK company limited by guarantee, does not provide services to clients.