Within the framework of the International Manufacturing Congress 2026 (IMC 2026), held on September 10 at the Vietnam Exhibition Center (VEC) in Hanoi, leading industry experts gathered for a high-level panel discussion titled “Smart factory at scale: Deploying AI, automation and digital technologies across Vietnam’s manufacturing sector.”
The session served as a dynamic platform for stakeholders to dissect pressing industry challenges, debate digital transformation strategies, and explore breakthrough operational solutions designed to propel Vietnam’s manufacturing landscape into its next phase of growth.
Reflecting on the prominent trends of core shifts in Vietnam's manufacturing sector, Dr. David Tw. Chia, Regional Managing Director, Beckhoff Automation Southeast Asia, noted that the biggest trend in recent years is not just AI but "Software-Defined Automation." This technology plays a smooth bridging role between hardware and software, addressing the core issue of scaling from the testing phase to mass production.
From an investment attraction perspective, Mr. Huynh Thien An, Director of LAPP Vietnam, pointed out that Vietnam is currently attracting large corporations not only due to labor costs or geographical location. The choice of global brands comes from the quality of the workforce, green energy infrastructure meeting 100% stringent standards, and the ability to master new technologies such as the semiconductor industry.
“In the next three years, competition will have to change. From low labor costs, we will shift to a higher labor and energy source, along with technological support. Especially, AI technology will help Vietnam become more competitive in the supply chain and attract FDI investment.
This shift also requires technological infrastructure to evolve. According to Mr. Doan Dai Phong, Deputy Chief Executive Officer, Viettel IDC, digital infrastructure and technology are no longer separate from the production process. In the next 3-5 years, individual smart factories will not be able to stand alone but must be deeply integrated into a tightly connected ecosystem with the entire value chain behind.
Agreeing with this trend of integration, Ms. Kimmy Nguyen, Regional Director of NTQ Europe, analyzed that while in the past 3-5 years, data primarily served internal management to optimize operations, in the next three years, factory data will become a direct competitive advantage of products, especially as customers increasingly care about traceability and quality reporting. “This will also be a competitive advantage when we know how to leverage factory data,” Ms. Kimmy Nguyen affirmed.
From a systemic perspective, Mr. Cao Dai An, Head of Technology and Solutions Office and AI Practice for ASEAN and Japan, Hitachi Digital Services, likened modernized data to the "brain" of enterprises. As integration deepens, data must be transparently linked globally to meet sustainable development standards. “We will see that data and digital workforce are among the trends in the coming years,” Mr. An remarked.
Resolving the infrastructure bottleneck: Synchronizing OT - IT and the clean data challenge
However, experts also noted that the gap between the idea of applying AI and its actual implementation in factories always contains many challenges. Mr.Huynh Thien An pointed out that the biggest barrier lies in the differences in protocols between devices and machinery imported from Japan, Europe, or the US. For AI to analyze and make accurate predictions, factories must solve the connectivity issue between OT (Operational Technology) data layers (devices, sensors, robots) and IT (office management systems) so that they can speak a common language.
Regarding infrastructure architecture, Mr. Đoan Dai Phong believes that for a factory, the core question is not about balancing edge computing or centralized computing, but rather where the workload needs to be processed.
For example, data from sensors or robots on the production line needs to be processed on-site to ensure fast response times. Conversely, image data from cameras in many factories needs to be centralized for analysis and training AI models.
A suitable technology architecture for factories must be a hybrid model, where the most important principle is to accurately determine the optimal processing location for each specific workload.
Adding to the role of technology, Mr. Cao Dai An stressed that technology is merely a supporting tool. In the testing phase, technology often takes center stage. However, when moving into actual operation, technology only plays the role of an activating factor for the transformation process, where data is the brain of the enterprise.
Especially, enterprises cannot apply pre-trained AI models based on general knowledge from around the world but must learn from their own clean data. “Industrial knowledge and internal data platforms are the core keys. If we only focus on individual technology projects, enterprises may succeed in the testing phase, but that data will not fully reflect operational realities,” Mr. An emphasized.
Therefore, the technology platform must be scalable enough to continuously update operational insights suitable for prioritized use cases. Concurrently, as AI and the digital workforce enter operations, enterprises must redesign their operational models and workflows. “From the railway manufacturing industry, transformers to high-tech robots, the question remains the same: how must processes and human roles change to develop in parallel with the digital workforce in the system,” Mr. An added.
Strategies to optimize technology investment for enterprises
Dr. David Tw. Chia pointed out the reality that many enterprises spend heavily on machinery, but these devices cannot communicate with each other, whether they are old or new machines. To address this, he recommends that factories should attach sensors to convert analog data to digital for old machines or leverage open protocols for new devices.
For leaders facing the challenge of limited capital but still wanting to create practical value, experts have proposed highly practical solutions.
Ms. Kimmy Nguyen suggested a strategy focusing on evaluating and collecting real-time data at a specific cluster of machines to address the downtime issue. Practical experience from electronic component projects shows that analyzing downtime causes by specific groups can help enterprises reduce downtime by up to 21% in just 6 months.
Discussing the digital transformation challenge for small and medium-sized enterprises (SMEs), Mr. Doan Dai Phong believes that these units should not build separate AI or digital infrastructure. According to him, self-investing in capital expenditure (CapEx) faces many significant barriers due to the complexity and short technology lifecycle, as AI chip lines often last only 2 to 3 years before becoming obsolete.
Therefore, Mr.Phong recommends that factories and SMEs should leverage shared infrastructure such as cloud computing or solutions from external providers. This approach helps enterprises shorten the time to market, optimize efficiency, and easily scale. Additionally, this expert also emphasized the importance of selecting the right business problem to implement in order of priority to achieve quick results.
Concluding the strategic discussion, Mr. Cao Dai An emphasized that leadership must plan a long-term roadmap of five to ten years suitable to the specific characteristics of the enterprise. Investing in a solid data platform from the outset, combined with addressing process bottlenecks, will be the key to helping enterprises maintain competitiveness and sustainable development amidst market fluctuations.
Google translate