National AI strategies are being written across the globe, but most governments are discovering that publishing a plan and actually building sovereign AI capability are two entirely different challenges. Three Southeast Asian countries that launched nearly identical AI roadmaps five years ago have since diverged sharply, revealing a pattern of failure that's now repeating in developed nations like the UK and Japan.

Why Do Similar AI Strategies Produce Such Different Results?

Indonesia, Malaysia, and Vietnam all entered the AI race around the same time with comparable ambitions. Indonesia published its National AI Strategy in 2020, Malaysia followed with its National Artificial Intelligence Roadmap in 2021, and Vietnam issued Decision 127 on AI research and development in January 2021. On paper, all three documents read like variations on the same script: priority sectors, talent pipelines, ethics principles, and promises of national competitiveness.

Five years later, the outcomes tell a starkly different story. Vietnam has passed a standalone, risk-based AI law effective from March 2026 and is moving toward explicit targets like contributing six percent of GDP through AI by 2030. Malaysia has become one of Southeast Asia's most aggressive magnets for data center investment, with 143 approved projects worth RM144.4 billion between 2021 and June 2025. Indonesia, the region's largest economy, is still waiting for a presidential signature on an umbrella regulation to guide AI adoption across ministries.

The real test of a national AI strategy doesn't come at the launch event. It comes in budget rooms, procurement decisions, institutional mandates, and the slow discovery of who actually owns the work after the cameras leave.

What Are the Three Most Common Reasons National AI Strategies Stall?

Experts have identified three recurring failure modes that explain why ambitious AI plans often collapse into inaction:

• The Aspiration-Execution Gap: Indonesia illustrates this most clearly. The country identified five priority sectors in 2020 (health, bureaucratic reform, education, food security, and mobility), but lacked a center strong enough to turn those ideas into actual direction. For years, Indonesia's AI agenda moved without a binding umbrella regulation, a dedicated implementing agency, or a delivery structure capable of disciplining sectoral fragmentation.
• Institutional Orphanhood: A strategy can create conversation, but execution requires an empowered owner. Indonesia's sectoral regulators have moved faster than the national framework meant to guide them. In April 2025, the Financial Services Authority launched AI Governance for Indonesian Banks, giving the banking sector clearer supervisory expectations than many other parts of the economy. The spokes are moving while the national umbrella still lags behind.
• Upstream Dependency: Malaysia shows this pattern most clearly. The country has attracted massive infrastructure investment, with 25 data center projects receiving Malaysia Digital status and expectations of 1,429 new jobs. But hosting data centers differs fundamentally from building domestic AI capability. The harder task is producing engineers who design systems, firms that build applications and models, and universities that train advanced talent at scale.

As one analysis noted, "a strategy can create conversation, but execution needs an empowered owner. A sovereign AI fund, fiscal incentives, and public-sector AI adoption may all help. Yet money only matters if it is attached to institutions that can spend it well".

How Is Vietnam Breaking the Pattern?

Vietnam's trajectory stands out as the most coherent of the three Southeast Asian nations. The country has moved from strategy to law, establishing a standalone, risk-based framework for AI providers and deployers. More importantly, Vietnam has built what experts call a "chain of command" around its AI ambitions.

Viettel, a state-linked conglomerate, has partnered with NVIDIA to develop a sovereign AI ecosystem and operates a cluster of 22 NVIDIA DGX B200 systems with reported performance of up to 1.5 exaFLOPs (a measure of computing speed). FPT has collaborated with NVIDIA on Vietnamese persona datasets and AI factory infrastructure. The result is alignment between legal direction, industrial champions, compute infrastructure, and national capability goals.

Vietnam has treated AI as an instrument of state capacity rather than only as a conventional technology sector. That gives Vietnam speed, but also a different vulnerability. Capability now runs through a small number of state-linked conglomerates and one dominant foreign chip ecosystem. The model depends on export licensing, vendor roadmaps, supply chains, and the balance sheets of a few national champions. Vietnam has not escaped dependency; it has simply organized it better.

Are Developed Nations Making the Same Mistakes?

The pattern isn't limited to Southeast Asia. The UK's Science, Innovation and Technology Committee recently warned that the government has "no coherent strategic framework" for how it will leverage its world-leading scientific research to advance wider diplomatic and economic goals.

Instead, the committee concluded that the UK takes an "opportunistic approach" to international agreements in science and technology. Without a clear framework backed by delivery plans, the UK risks "substituting activity for strategy," weakening its international credibility and wider ambitions.

"The UK is in the premier division of science and the premier division for diplomacy, but we don't know where we stand in the field of science diplomacy. As geopolitics is turned upside down and the world becomes increasingly competitive, we must be able to leverage our world-class science and research to advance our diplomatic and economic goals. Without a clear plan, the government will be unable to achieve this," stated Dame Chi Onwurah, Chair of the Science, Innovation and Technology Committee.

Dame Chi Onwurah, Chair of the Science, Innovation and Technology Committee

The committee warned that the US's recent restrictions on some AI models highlights the risk of relying on allies for access to technologies critical to economic growth and national security. The UK must protect its tech sovereignty by setting out realistic ambitions for sovereign capabilities in key sectors like AI, quantum, and space.

What Is Japan's Approach to Breaking the Cycle?

Japan is attempting to address the execution problem head-on by establishing a new National AI Reform Council. The council, announced within the context of government guidelines for 2026, will work on a legal and regulatory framework aimed at reforming the existing structure regarding AI technologies in the country.

The new committee will replace the current Digital Administrative and Fiscal Reform Council and will develop policy aimed at boosting AI innovation in view of Japan's demographic and economic challenges. Japan has gradually invested in AI through programs on sovereign large language models (LLMs, or AI systems trained on vast amounts of text), cutting-edge semiconductor production, robots, cloud infrastructure, and the use of AI in providing public services.

The establishment of the AI reform council represents Japan's recognition that outdated legislation is among the most formidable obstacles to widespread AI acceptance. Current laws and administrative procedures need to change to accommodate AI-based models of doing business and provision of services, as well as automated decision-making.

Japan's strategy stems from the realization that AI use in governance must be enhanced to boost efficiency, reduce costs, and provide better services to citizens. With an aging population and dwindling labor force, the government believes AI has become necessary. Rather than eliminating jobs, AI technologies are intended to enhance human abilities through automation of routine tasks and improved decision-making processes.

How to Build a Sovereign AI Strategy That Actually Works

• Establish Clear Institutional Ownership: Designate a single empowered agency or council with authority over AI policy, budget allocation, and sectoral coordination. Indonesia's failure to do this allowed sectoral regulators to move faster than the national framework, creating fragmentation rather than alignment.
• Align Legal Frameworks with Industrial Champions: Create binding regulations that work in concert with domestic firms and state-linked entities. Vietnam's approach of pairing its AI law with partnerships between Viettel, FPT, and NVIDIA created a coherent chain of command from policy to execution.
• Modernize Regulations Alongside Infrastructure Investment: Don't wait for perfect laws before building compute capacity, but don't build infrastructure without regulatory clarity either. Japan's AI Reform Council is addressing this by updating laws in parallel with infrastructure development, reducing the aspiration-execution gap.
• Build Domestic Capability, Not Just Infrastructure: Attracting data centers and cloud investment is an achievement, but it's not the same as building the engineers, firms, and universities needed to design and deploy AI systems domestically. Malaysia has solved the infrastructure problem but still faces the harder task of deepening capability.
• Create Accountability Structures Beyond Government and Market: Universities, research institutions, and civil society need enough capacity to scrutinize both state and market actors. A multiplex digital ecosystem requires a third force beyond government direction and market supply.

The deeper issue underlying all these failures is that most national AI strategies are still written as state-market bargains, where the state sets direction and the market supplies capability. But AI now operates in what scholars call a "multiplex world": many actors, no single director, and power distributed across institutions, firms, infrastructures, standards, and publics.

A serious strategy for a multiplex digital ecosystem needs more than ambition and infrastructure investment. It needs institutional clarity about who owns the work, legal frameworks that enable rather than obstruct innovation, and accountability structures that can scrutinize both government and private sector actors. Without these elements, even the most carefully written AI strategy will stall in the same way Indonesia's has, or produce infrastructure without capability like Malaysia's, or concentrate risk in a small number of state-linked firms like Vietnam's.

The countries that succeed in building genuine sovereign AI capabilities won't be those with the most ambitious documents. They'll be the ones that answer the unglamorous question first: who will have the authority, resources, and public accountability to make this real after the launch event is over?