Vietnam can build AI sovereignty based on specialized systems designed to align with national strengths. (Source: Magnific)

As countries worldwide accelerate AI development and secure semiconductor supply chains, Vietnam faces critical strategic choices on how to build its own AI ecosystem amid an increasingly fragmented global technology landscape.

Accordingly, Vietnam should deprioritize building large-scale, multi-functional AI models and instead focus on developing specialized AI systems based on Vietnam's high-quality data sources.

Attempting to build a general-purpose, multi-functional AI system from scratch to compete with global tech giants could cost billions of dollars, enormous energy resources, and hardware access. Instead of following strategies from the U.S. or China, Vietnam needs to pursue a more focused path based on its unique advantages.

Vietnam should not enter a race by emulating what major tech corporations are doing. Rather than wasting resources on large-scale and costly AI models, Vietnam's priority must be specialization. Vietnam should concentrate all resources on building specialized AI infrastructure based on high-quality Vietnamese datasets. This could be the most practical path for Vietnam to achieve AI sovereignty.

Specialized AI systems can support areas where Vietnam has data and expertise advantages. By integrating local data into specialized AI infrastructure, we can build world-class AI tools specifically designed for agriculture, regional healthcare, or Vietnam's legal and administrative systems. This approach will help Vietnam develop AI tools that are cost-effective, highly efficient, and achieve technological autonomy.

Dr. James Kang is currently a senior lecturer in Computer Science at RMIT University Vietnam. (Source: RMIT)

A strong legal framework needed

It's not difficult to recognize that U.S. export restrictions on advanced AI chips, including NVIDIA H200, have accelerated the technological decoupling between the U.S. and China.

NVIDIA is currently in fierce competition to secure enough semiconductor components from key suppliers like Samsung to meet soaring global demand. As major tech corporations have secured supply chains, it is unrealistic for Vietnam to easily acquire sufficient hardware to build large-scale AI infrastructure.

Meanwhile, some recent advancements in Chinese AI heavily rely on "model distillation" techniques, which involve using advanced U.S. AI systems to train their models.

This approach saves China time but also means Beijing will always lag behind, as U.S. technology continues to advance. Another issue is the lower quality of data China can access compared to the U.S., limiting Chinese AI's appeal to global users.

While acknowledging concerns about balancing innovation and governance, Vietnam has its own institutional advantages in developing autonomous AI.

When it comes to data protection while maintaining an open economy for domestic businesses, Vietnam indeed has a significant structural advantage, allowing for the development and implementation of long-term national AI policies with consistency and speed.

Accordingly, the Personal Data Protection Law and the Artificial Intelligence Law, which came into effect this year, have laid a crucial foundation for protecting the nation's sensitive data while supporting domestic innovation. These two legal frameworks combined will aid Vietnam in tightly managing data flow and sharing.

Simultaneously, policymakers can foster innovation by establishing legal testing mechanisms for domestic businesses. Specifically, the government can use centralized governance to establish a flexible legal testing mechanism – a testing environment with streamlined procedures and safe zone provisions.

Looking to the future, Physical AI could become Vietnam's next major opportunity. Through close collaboration with tech companies heavily investing in smart hardware and robotics, Vietnam can leverage existing technology partnerships to develop capabilities beyond manufacturing. Since Physical AI requires expertise in both software and hardware, this could be the path for Vietnam to learn, innovate, and compete in the next phase of the global AI economy.