In March 2026, Vietnam officially began the implementation of Law 134/2025/QH15, a significant state-led effort to regulate the burgeoning artificial intelligence sector. This top-down legislative framework represents one of the most assertive moves by a Southeast Asian government to manage AI technology. By establishing this law, Hanoi intends to bolster sovereign control over data flows, increase transparency, and mitigate the diverse risks associated with AI deployment. The move sets a distinct benchmark for the region, which has traditionally leaned toward voluntary compliance and soft-law guidelines, such as those promoted by ASEAN.
The law introduces a comprehensive three-tier risk management system—categorizing AI activities as high, medium, or low risk—with regulatory oversight scaling according to the level of risk identified. These rules apply to both domestic and foreign entities operating within Vietnam. To foster development, the government has included a trial phase that allows tech developers to test evolving AI models without facing strict penalties, aiming to balance necessary regulation with the country's broader socio-economic and high-tech ambitions.
However, the legislation faces criticism for its potential disconnect from broader regional and environmental realities. By moving toward a standalone, forceful legal framework, Vietnam risks creating friction with ASEAN’s existing voluntary approach. Observers note that this could trigger a high-tech race to the bottom, as venture capital often gravitates toward jurisdictions with fewer regulatory burdens. ASEAN’s own Guide on AI Governance and Ethics, first issued in February 2024 and expanded in January 2025 to include generative AI, emphasizes principles like fairness, privacy, and safety. There is speculation that ASEAN could shift toward a legally binding framework by the end of 2026 via the ASEAN Digital Economy Framework Agreement.
Furthermore, the law is notable for its silence on environmental degradation and natural resource allocation. This oversight is particularly critical given the heavy resource demands of AI infrastructure. According to the World Economic Forum, a standard 1 megawatt (MW) data center can consume 25.5 million liters of water annually—a volume equivalent to the daily water usage of 300,000 residents in a medium-sized Vietnamese city. As the World Bank identifies Vietnam as highly susceptible to climate change and rising coastal waters, the diversion of water for cooling data centers poses a significant threat to local and regional reserves.
Ultimately, the effectiveness of Vietnam’s new AI framework remains in question. Experts argue that without addressing the governance of essential commodities like land and labor, and failing to harmonize regulatory coordination, the law may inadvertently exacerbate economic dislocation and social instability. As global powers like the United States, China, Russia, and the European Union continue to treat AI as a central pillar of their strategic grand designs, Vietnam’s ability to successfully navigate these complexities will likely influence its long-term regional standing.
