Distributed AI in Vietnam: from ambition to leadership

Vietnam has given itself four years to become a top 4 ASEAN and top 50 global AI power. That is an aggressive timeline for any country, but Vietnam is not starting from a standing start: 74% of its businesses are already using AI, up 39% in a single year. The ambition is real, and so is the pace of adoption.

The problem is capacity. Our new report, From Cloud to Edge: The Value of Distributed AI in Vietnam, finds that demand for AI inference compute in the country could grow 30 times by 2030. Meeting that from data centres alone would mean multiplying planned capacity sixfold, and even then, Vietnam would likely face an 85% shortfall in inferencing capacity. Cloud cannot get Vietnam to its target on its own.

The case for splitting the workload

Distributed AI, processing split across cloud, edge and device rather than routed entirely to remote data centres, offers a way around that ceiling. It is also a better fit for the kind of AI Vietnam is trying to scale. Decision 21/2026/QD-TTg already names edge-AI cameras, autonomous mobile robots and UAVs as strategic technology products, and applications like these depend on qualities cloud struggles to deliver: instant response for robots working alongside people, resilience for drones in disaster response, and the ability to keep sensitive data onshore in support of Vietnam’s digital sovereignty goals. Distributed AI also eases pressure on the grid, processing locally rather than in energy-intensive data centres, which matters given Resolution 70’s energy security mandate.

What it’s worth

We estimate distributed AI could generate up to USD 30 billion in annual economic value for Vietnam by 2035, concentrated in three areas:

• Autonomous manufacturing and robotics (over USD 13 billion), through higher labour productivity, less downtime and fewer safety incidents in a sector worth roughly a quarter of GDP and around USD 400 billion in exports.
• Smart city management (over USD 11 billion), from reduced traffic congestion, lower crime and more efficient power grids.
• AI-enhanced healthcare delivery (over USD 6.1 billion), through delivery efficiencies, fewer hospitalisations and fewer sick days.

The cost of getting there

None of this happens by default. We estimate it will take around USD 5.2 billion of incremental investment between 2026 and 2035, roughly 14% of Vietnam’s relevant science and technology budget for 2026, for annual returns we project at more than five times that outlay.

Vietnam already has more of the foundation in place than most countries at this stage: an AI Law in force since March 2026 with a risk-based, tiered governance structure, tax incentives for digital technology projects, and 5G coverage that already reaches 90% of outdoor areas, with 6G research underway.

What is missing is coordination. Our report sets out four policy priorities to close that gap:

1. Leverage Vietnam’s role as APEC 2027 host to champion AI standards and interoperability across the region.
2. Establish data portability and vendor diversity conditions on the national AI Data Centre programme to keep the ecosystem open.
3. Front-load 5G and 6G investment so networks keep pace with agentic and physical AI.
4. Embed concrete benchmarks for distributed and edge AI into Resolution 57, the Law on AI, and the National Data Strategy.

Vietnam has set itself a genuinely fast timeline. Distributed AI is what makes hitting it plausible, and the choices made over the next few years, particularly with the world watching at APEC 2027, will decide how much of that US$30 billion opportunity Vietnam actually captures.