AI is now embedded in most corporate digital transformation strategies. Yet many businesses continue to face significant barriers to implementation due to high investment costs, complex infrastructure requirements and limited access to computing resources.

From digital transformation to national digital capability

Speaking at the Vietnam National Digital Forum (DIGIFEST) 2026, Deputy Minister of Science and Technology Bui Hoang Phuong stressed that digital transformation is not simply about investing in infrastructure, software or technology devices. More importantly, it must create practical value for citizens and businesses while improving operational efficiency.

"The essence of digital transformation is not purchasing technology. It must begin with solving real-world problems faced daily by local authorities, businesses and citizens. The goal is not to perform existing tasks in a digital environment, but to do them better, faster, simpler, more transparently and more effectively," Phuong said.

According to Tran Duy Ninh, Director General of the National Digital Transformation Agency, national digital capability is the ability to mobilize, organize, control and effectively utilize digital resources to create value for socio-economic development. In this framework, the highest layer is not infrastructure or data, but the ability to master strategic technologies - a factor that will determine a country's future competitiveness.

Pham Quang Hieu, Director General of the Information Technology Department under the Ministry of Justice, highlighted the importance of developing a specialized large language model for Vietnamese legal services.

"One of our key priorities is building a large language model tailored to Vietnamese law. Developing domestic AI models helps safeguard data sovereignty, digital sovereignty and national information security. On that foundation, we will develop AI assistants to support legislative drafting, legal review processes and easier public access to legal information," Hieu said.

Mastering AI is about more than data

The rapid rise of AI is changing how nations define technological capability. In the past, storing data domestically was often viewed as the cornerstone of digital sovereignty. In the AI era, however, data is only one piece of the equation.

Experts argue that sovereign AI is not merely about keeping data within national borders. It also involves controlling the entire AI value chain, including computing infrastructure, datasets, AI models, engineering talent and the ability to operate and advance the technology.

If data is the raw material of the digital economy, computing power is the factory of the AI economy. A country may possess vast amounts of data, but without the ability to process, train and deploy AI models at scale, it will struggle to generate meaningful value.

As demand for AI applications continues to rise, access to computing infrastructure is becoming one of the biggest obstacles facing businesses. In Vietnam, demand for AI infrastructure is already growing rapidly.

Against this backdrop, a growing number of domestic technology companies have begun investing in AI capabilities as a long-term strategic priority. Several Vietnamese firms are pursuing AI development across three layers - infrastructure, platforms and applications - with the goal of building sustainable AI capacity rather than isolated AI features.

Solving the biggest barriers to AI deployment

Speaking at DIGIFEST 2026, Vu Thanh Tung, Chief Executive Officer of GreenNode, said most businesses recognize the opportunities created by AI but continue to face significant implementation challenges.

"The needs of Vietnamese enterprises are not very different from the systems we currently operate. Whether it is e-commerce, digital wallets, online platforms or AI applications, all require infrastructure that is powerful, reliable and scalable," Tung said.

According to Tung, a single AI server system can cost tens of billions of Vietnamese dong, equivalent to hundreds of thousands of US dollars. Yet many businesses have not fully defined the problems they want AI to solve or how they will measure returns on investment.

Investment costs are only part of the challenge. Companies must also address issues involving data centers, electricity consumption, network connectivity, storage systems, cybersecurity and highly skilled operations teams.

Data sovereignty and regulatory compliance present another major concern. When data and AI infrastructure are hosted entirely overseas, businesses face additional challenges related to security, data governance and compliance requirements, particularly in sectors such as finance, banking, insurance and public services.

AI is already delivering measurable value across industries by solving specific operational challenges. For example, GreenNode's Intelligent Document Processing (IDP) solution has helped financial institutions, banks and insurance companies process hundreds of thousands of documents while saving tens of billions of Vietnamese dong.

"We focus on developing AI platforms and solving real-world problems for customers. What matters is not the technology itself, but the value it creates. AI only generates economic value when businesses can use it to improve productivity and operational efficiency," Tung said.

In the long run, sovereign AI will not be measured solely by where data is stored, but by a country's ability to own the infrastructure, technology and human resources needed to transform data into value. This will be essential if Vietnam is to move beyond being a technology adopter and become a meaningful contributor to the development of future strategic technologies.

Thai Khang