Vietnam enacted its Law on Artificial Intelligence on 10 December 2025. It took effect on 1 March 2026. In just over three months, the country went from having no dedicated AI legislation to operating one of the most comprehensive AI governance frameworks in the world — a 36-article law covering risk classification, transparency obligations, incident management, and enforcement mechanisms including fines of up to 2% of annual revenue.
This isn’t a statement of intent. It’s binding law, and it applies to every organisation involved in AI research, development, deployment, or use within Vietnam’s borders, including foreign providers. If you’re operating AI systems in the Vietnamese market — or planning to — here’s what you need to know.
How does Vietnam classify AI systems?
The law introduces a three-tier risk classification: low, medium, and high. AI providers must self-classify their systems before deployment and notify the Ministry of Science and Technology for medium- and high-risk systems through a national one-stop AI portal. Low-risk disclosures are voluntary.
High-risk AI systems — defined as those posing serious risks to human health, fundamental rights, public interests, or social order — face the most stringent requirements: data security protocols, impact assessments, human oversight mechanisms, and ongoing monitoring obligations. The Prime Minister’s office will publish a specific list of high-risk AI system categories, with draft versions expected for public comment in the coming months.
If an AI system is modified in ways that increase its risk level, providers must coordinate reclassification with the regulator. The governance model parallels the EU AI Act’s provider-deployer framework, with one notable difference: Vietnam separates the “developer” role (design and training) from the “provider” role (market placement), exempting non-market research and development from most compliance obligations to protect innovation.
What do foreign AI providers need to do?
Foreign providers deploying AI systems in the Vietnamese market must appoint a legal representative in Vietnam — a clear point of accountability for regulatory authorities. This requirement applies regardless of whether the provider has a physical office in the country.
AI systems operating in Vietnam before the law’s effective date have 12 months (until March 2027) to align with the new compliance requirements. Legacy systems in healthcare, education, and finance have 18 months. During the transition period, existing systems can continue to operate normally unless the competent authority identifies substantial risks.
The maximum administrative fine for violations is VNĐ 2 billion (approximately $76,000) for organisations and VNĐ 1 billion for individuals. For serious violations, fines can reach up to 2% of the violating organisation’s revenue in the preceding year. The revenue-based fine structure mirrors the approach taken by the EU’s GDPR and AI Act.
What support does Vietnam offer for AI development?
The law isn’t purely regulatory — it includes significant incentive mechanisms. AI-related activities receive the highest tier of investment incentives available under Vietnamese law.
The National AI Development Fund offers grants, loans, and preferential financing to domestic startups, SMEs, and foreign investors developing AI capabilities in Vietnam. A voucher scheme allows startups training Vietnamese-language models or working with Vietnamese data to access high-performance computing services — cloud GPUs from Viettel or VNPT — at reduced cost. This directly addresses one of the biggest barriers to AI development in emerging markets: the cost of compute.
Regulatory sandbox mechanisms allow selected AI applications to be tested under relaxed compliance conditions with fast-track assessment processes. The government has explicitly framed this as a tool to lower market entry barriers while supporting responsible innovation.
How does this compare to other Asian AI frameworks?
Vietnam’s approach sits between Japan’s obliquely promotional framework (the AI Promotion Act imposes no penalties for non-compliance) and China’s more interventionist model (which combines state-backed open-source promotion with direct technology policy).
The closest structural comparison is the EU AI Act, and Vietnam’s law explicitly draws on European thinking — particularly around risk-based classification, transparency requirements, and provider obligations. This makes sense strategically. Vietnam is positioning itself to attract both EU and US investment by demonstrating regulatory predictability that mirrors Western governance expectations while offering Asian-market access and cost advantages.
South Korea’s Framework AI Act shares similar ambitions around comprehensive governance, but Korea’s implementation runs through a multi-year roadmap (2026–2028) with 99 execution tasks. Vietnam has moved faster and with less institutional complexity — partly because the country’s National Assembly has demonstrated a willingness to pass technology legislation at unusual speed (the companion Digital Technology Industry Law was enacted in June 2025, just months before the AI Law).
What should companies do now?
For companies already operating AI systems in Vietnam, the immediate priorities are classification and gap analysis. Self-classify every AI system against the three-tier framework. Identify which systems are likely to fall into medium or high-risk categories. Begin preparing the notification documentation for the national AI portal.
Foreign providers without a Vietnamese legal representative need to establish one before enforcement begins to bite. The 12-month grace period for existing systems means March 2027 is the hard deadline — but the Ministry of Science and Technology’s implementing decrees, expected throughout 2026, will clarify specific requirements.
For companies considering the Vietnamese market, the AI Law should be read as a signal of regulatory maturity rather than a barrier to entry. The incentive package — fund access, voucher schemes, sandbox mechanisms — is designed to attract AI investment. But the compliance architecture is real, and companies entering the market after March 2026 don’t get the transition period.
Vietnam’s IT market is one of Asia’s most dynamic — the government targets 150,000 digital technology firms by 2035 and the digital economy reaching 30% of GDP by 2030. The AI Law is part of a broader legislative push that includes the Digital Technology Industry Law and national strategies for semiconductors and digital infrastructure. Companies that engage with the regulatory framework early will be better positioned than those that treat compliance as an afterthought.
Read the full AI Ecosystem Across Asia 2026 report — market data, country analysis, company profiles, investment landscape, and 2026–2030 outlook → digitalinasia.com/reports
