The Decision Is About More Than the Model
When an organisation selects AI, the location of data processing, ownership of infrastructure and terms of service all affect cost, continuity and governance. These questions are bringing Sovereign AI into government and business decisions.
Deloitte’s latest report states that global investment in Sovereign AI is expected to surpass $100 billion in 2026, more than tripling from 2024.
What Is Sovereign AI?
Sovereign AI describes the ability of a country or organisation to decide who controls its infrastructure, data and AI models. Discussions usually cover three dimensions:
- Compute Sovereignty: domestic datacenters and GPU/AI chip capacity
- Data Sovereignty: rules for where data is stored, processed and accessed
- Model Sovereignty: the ability to develop and maintain models for local language and context
Why 77% of Business Leaders Say “Location” Matters
Deloitte’s survey found that 77% of business leaders say the location where AI is developed and deployed affects provider selection. Organisations should examine:
- Data law and contract terms: where data goes, who can access it and which requirements apply
- Cross-border access risk: overseas processing may introduce more parties and rules to review
- Business continuity: trade restrictions or service changes may affect access to chips, models or APIs
Developments Across ASEAN
Several ASEAN countries are investing in AI infrastructure and policy:
- Singapore: investing more than $1B in National AI Strategy 2.0
- Indonesia: building a sovereign cloud for government AI
- Vietnam: introducing ASEAN’s first AI law
- Malaysia: attracting Nvidia and Microsoft to build an AI hub
Where Thailand Stands
Thailand has several related infrastructure and policy developments:
- 5GW datacenter buildout: large-scale projects that would add compute capacity in Thailand
- Google, AWS, Microsoft have announced cloud region investments in Thailand
- Thailand’s draft AI law is currently under review
Questions remain about AI skills, energy supply for datacenters and regulations that are still developing.
Implications for Thai Organizations
Map the data flow. Organisations using foreign AI APIs should document which data is sent where, how long providers retain it and which processors are involved before assessing PDPA and contractual requirements.
Assess portability. If the organisation wants hybrid deployment (cloud + on-premise), review how data, models and workflows can be moved, as well as the skills and cost required to operate them.
Set selection criteria first. Business, legal, security and technology teams should agree which work must be processed in Thailand, which work can use overseas services and why.
Sources:
