Nguyen Van Minh, founding head of Hardware Efficient AI and AI Native Wireless at Huawei's Research Centre in France, told Vietnam News Agency that Vietnam's semiconductor strategy should move from planning to execution.
Minh said Vietnam and France have no joint programmes in core semiconductor technologies yet. Cooperation between Vietnam's National Innovation Centre and Dassault Systèmes supports industrial digitalisation, digital twins and product lifecycle management, but has not reached chip design, AI accelerator architecture, semiconductor intellectual property or advanced packaging.
He argued that Vietnam should view semiconductors beyond 2 nanometre and 3 nanometre fabrication. The market includes power chips, communication chips, sensors, memory, industrial processors, automotive chips, IoT components and edge AI hardware.
His suggested path starts with semiconductor materials, wafers and upstream supply chains, then advanced packaging, testing, chiplets and heterogeneous integration. Mature process technologies such as 28 nanometre nodes could serve industrial, automotive, IoT and edge AI demand before Vietnam attempts higher cost fabrication targets.
For AI, Minh pointed to applications in manufacturing, agriculture, healthcare, transport, education and public services, backed by local data and domain knowledge. He also urged Vietnam to build capabilities in algorithm optimisation, computing architecture, AI accelerators, hardware design and energy efficient computing.
The talent gap remains central. Minh said Vietnam needs skills across algorithms, systems, hardware architecture, chip design, manufacturing, packaging and testing.
The-Delta's Takes Editorial Analysis
What This Means
For Vietnamese ministries, the decision path is sequencing. A credible semiconductor programme needs target product categories, shared technical facilities, procurement demand, and accountable owners for each stage of the value chain. Advanced packaging, chiplet work, testing, mature process design, AI accelerator architecture and edge AI hardware offer a more reachable operating path than chasing the most expensive fabrication nodes first.
For Vietnamese universities and engineering firms, the message is that talent policy must follow product architecture. Training thousands of generic semiconductor workers will produce weak output if courses are disconnected from specific needs in system design, verification, packaging, test, power electronics, memory, communication chips and AI hardware. Foreign partners should also read this as a filter. Vietnam will need partners who transfer operating methods, technical standards and applied project discipline, not only seminar level cooperation.
Execution Reality Medium
Vietnam now has policy attention, foreign investor interest and a growing talent pipeline. The bottleneck sits in programme design. A working execution chain needs budgets, laboratories, chip design tools, packaging and test access, industry mentors, university pipelines, and customers willing to buy early Vietnamese products. France can help with research institutions, industrial modelling and technical training, but the article notes no core Vietnam France semiconductor programme yet. The practical test is whether Vietnam can turn bilateral interest into funded work packages tied to product categories and named institutions.
The Signal
Vietnam's semiconductor policy is moving from ambition to prioritisation. The strongest opportunity may sit in the middle layers of the value chain where technical depth, application knowledge and production discipline matter more than owning the most advanced fabrication process. That shift would align Vietnam with industrial demand in automotive, IoT, edge AI, communications and power electronics. It also reframes foreign cooperation. The useful partners will be those who help Vietnam build repeatable execution capacity across design, packaging, testing and applied AI hardware.