AI Infrastructure and ASEAN's Sovereignty

Global enterprise spending on cloud infrastructure continues its unabated climb, recently hitting $143 billion, according to a report from Data Center Dynamics. This eleventh straight quarter of growth is propelled by a worldwide race to build the foundational capacity for artificial intelligence. While much attention is paid to the AI models themselves, the underlying story is one of massive capital expenditure on the physical hardware and data centers that give them life. For ASEAN, this global trend presents a distinct and pressing set of questions about its economic future.

The Infrastructure Imperative

The development of artificial intelligence is fundamentally a story about infrastructure. The computational power required to train and operate sophisticated AI systems necessitates vast, power-hungry data centers. The global surge in spending reflects a concentration of this capacity in the hands of a few large-scale cloud providers. As nations and corporations alike rush to adopt AI, they become increasingly dependent on this foundational layer of technology.

ASEAN member states are not bystanders. Governments and enterprises across the region are eager to integrate AI to boost productivity and create new economic avenues. However, the region largely relies on external providers for this critical infrastructure. This creates a long-term dependency that extends beyond simple commercial relationships, touching on economic resilience and national autonomy.

The Sovereignty Question

The true competition in the AI era is not about creating a national alternative to every popular model. As the book "ASEAN Rising" outlines, the more substantive issue is one of control over the core components of the digital economy. The book notes that "Sovereign AI in ASEAN will be tested not by model launches but by who controls the compute, the data layer and the digital identity rails that sit underneath everyday economic life."

This reframes the challenge for ASEAN. The first element is compute-the physical data centers and processing hardware. The second is the data layer-the aggregation and governance of the public and private data that fuels AI systems. The third is the digital identity rails-the systems that verify and manage the identities of individuals and entities transacting in the digital economy. Without a degree of sovereign control over these three layers, any national AI strategy rests on a weak foundation, subject to the priorities and pressures of external technology providers.

Execution and Trust

Building this sovereign capacity is a monumental task of execution and capital allocation. It involves creating policy frameworks that encourage investment in local data center infrastructure, as seen in countries like Thailand, Malaysia, and Indonesia, which are attracting significant interest from hyperscalers. It also requires a trusted institutional ability to manage national data and identity systems securely and effectively.

This is not a call for technological isolationism. Rather, it is about building the domestic institutional and infrastructural muscle required to engage with global technology partners on more equal terms. It involves difficult policy choices about data governance, cross-border data flows, and the role of state-owned enterprises versus private capital in developing these foundational assets. The goal is to ensure that the benefits of the AI economy are not just consumed in ASEAN, but are also built and substantially managed from within it.

What to watch

The key indicators of ASEAN's progress will not be found in announcements of new AI models. Instead, watch the flow of foreign and domestic investment into data center construction and renewable energy projects to power them. Monitor the progress of national digital identity programs and their interoperability across the region. Pay close attention to the substance of data governance laws and how they balance the need for privacy and security with the economic imperative of putting data to productive use.