Vietnam’s semiconductor push should prioritize packaging, testing, mature-node chips and AI hardware before attempting to compete in leading-edge manufacturing, according to Dr. Nguyen Van Minh, founding head of Hardware-Efficient AI & AI-Native Wireless at Huawei’s Research Centre in France.
Speaking with the Vietnam News Agency in Paris, Minh argued that Vietnam’s biggest challenge is no longer drawing up a semiconductor strategy. It is turning one into funded programs, shared technical infrastructure, industry-university research links and products that can actually find a market.
Minh’s advice is a useful corrective to the global tendency to equate semiconductor success with 2nm and 3nm logic chips. Vietnam has opportunities across power semiconductors, communications ICs, sensors, memory, industrial processors and components for automotive, IoT and edge-AI equipment — markets where mature processes and integration skills can be commercially significant.
His proposed sequencing is equally pragmatic: develop semiconductor materials and wafer supply-chain capabilities, then expand advanced packaging, testing, chiplets and heterogeneous integration. Only after building those foundations should Vietnam progressively target mature manufacturing nodes such as 28nm.
For Windows and enterprise IT readers, that approach aligns with where hardware demand is broadest. Edge systems in factories, transport, public services and retail do not necessarily need the most advanced processor node; they need dependable, power-efficient silicon with long product lifecycles and a supportable supply chain.
That distinction matters. Dassault Systèmes and NIC have already positioned their work around workforce development and semiconductor ecosystem building, while Vietnam has established a national chip-prototyping support centre and adopted policy targets for advanced packaging and testing. But software platforms, simulation and digital-twin expertise are enabling layers, not substitutes for a local chip-development pipeline.
The missing step is a tighter connection between those tools and concrete engineering output: reusable semiconductor IP, validation capacity, package design, prototype runs and customers willing to deploy Vietnamese-designed components.
Vietnam’s policy ambition includes training 50,000 semiconductor engineers by 2030, but quantity alone will not establish an ecosystem. The country must decide where it wants technological autonomy, where it wants export-scale commercial competition, and which chip categories can plausibly meet both goals.
The near-term test is whether Vietnam can turn its strategy into projects with named owners, equipment, engineering talent and procurement demand. In semiconductors, the plan is the easy part; the production and integration capacity is the asset that compounds.
The 2nm race is not Vietnam’s immediate contest
Minh’s advice is a useful corrective to the global tendency to equate semiconductor success with 2nm and 3nm logic chips. Vietnam has opportunities across power semiconductors, communications ICs, sensors, memory, industrial processors and components for automotive, IoT and edge-AI equipment — markets where mature processes and integration skills can be commercially significant.His proposed sequencing is equally pragmatic: develop semiconductor materials and wafer supply-chain capabilities, then expand advanced packaging, testing, chiplets and heterogeneous integration. Only after building those foundations should Vietnam progressively target mature manufacturing nodes such as 28nm.
For Windows and enterprise IT readers, that approach aligns with where hardware demand is broadest. Edge systems in factories, transport, public services and retail do not necessarily need the most advanced processor node; they need dependable, power-efficient silicon with long product lifecycles and a supportable supply chain.
France links digital twins to a larger industrial goal
Minh said cooperation between Vietnam’s National Innovation Centre and France’s Dassault Systèmes has helped advance industrial digital transformation, digital twins and product lifecycle management, but has not yet moved into core semiconductor capabilities such as chip design, IP, accelerator architecture or advanced packaging.That distinction matters. Dassault Systèmes and NIC have already positioned their work around workforce development and semiconductor ecosystem building, while Vietnam has established a national chip-prototyping support centre and adopted policy targets for advanced packaging and testing. But software platforms, simulation and digital-twin expertise are enabling layers, not substitutes for a local chip-development pipeline.
The missing step is a tighter connection between those tools and concrete engineering output: reusable semiconductor IP, validation capacity, package design, prototype runs and customers willing to deploy Vietnamese-designed components.
Talent and execution become the real constraints
Minh emphasized that Vietnam needs specialists across the full value chain, from algorithms and system engineering through hardware architecture and IC design to manufacturing, packaging and testing. That requirement is substantially wider than simply training chip designers — and it is likely to be harder to satisfy.Vietnam’s policy ambition includes training 50,000 semiconductor engineers by 2030, but quantity alone will not establish an ecosystem. The country must decide where it wants technological autonomy, where it wants export-scale commercial competition, and which chip categories can plausibly meet both goals.
The near-term test is whether Vietnam can turn its strategy into projects with named owners, equipment, engineering talent and procurement demand. In semiconductors, the plan is the easy part; the production and integration capacity is the asset that compounds.