Cloud infrastructure: the foundation of digital government in the AI era

September 22, 2026 - 10:46
The policy shift marks a significant step towards the country's goal of ranking among the world's top 50 in the E-Government Development Index (EGDI) by 2030, positioning cloud as the foundation of next-generation digital government.
Cloud infrastructure has been elevated from a project-level option to a core principle of Việt Nam's digital architecture. — Photo courtesy of the Ministry of Science and Technology

HÀ NỘI — Cloud infrastructure has been elevated from a project-level option to a core principle of Việt Nam's digital architecture under three strategic government decisions on digital transformation, the national digital architecture and artificial intelligence (AI).

The policy shift marks a significant step towards the country's goal of ranking among the world's top 50 in the E-Government Development Index (EGDI) by 2030, positioning cloud as the foundation of next-generation digital government.

From project-based infrastructure to shared architecture

Decision No. 1266/QĐ-TTg, dated July 14, approving the National Digital Transformation Strategy for 2026–2030, stipulates that digital infrastructure must become 'an essential and strategic national infrastructure', serving as the foundation for data, artificial intelligence (AI), digital government, the digital economy and the digital society.

For the public sector, digital infrastructure is envisioned as a unified, flexible and scalable system capable of supporting operations from the central government to local authorities.

Decision No. 1425/QĐ-TTg, dated July 29, approving Version 1.0 of the National Digital Architecture Framework, serves as the technical blueprint for implementing the National Digital Transformation Strategy.

The framework organises shared national digital capabilities into four layers: digital infrastructure and cybersecurity; data and core platforms; shared applications and business services; and interaction channels and performance measurement. Digital infrastructure forms the foundational layer underpinning the entire architecture.

The decision explicitly adopts the principles of 'Cloud First' – prioritising the use of cloud infrastructure and platforms – and 'shared by default', giving priority to the reuse of existing digital infrastructure over fragmented, standalone investments.

Under these decisions, cloud is no longer merely a technology choice for individual projects but a national design principle, underpinning the country's digital architecture alongside decentralisation and the delegation of authority.

Under this model, the central government is responsible for developing and operating shared digital platforms, while ministries, sectors and local authorities connect to and build upon them. Government agencies and provincial administrations have six months from the issuance of the framework to complete their own digital architecture, with a deadline of the end of January 2027.

This principle also reshapes the public investment model. Instead of purchasing and maintaining dedicated servers for individual projects, government agencies are gradually shifting to cloud infrastructure as a service, paying according to actual usage.

The approach enables faster application deployment while reducing the burden of managing the full lifecycle of hardware and related infrastructure.

The National Data Centre has been operational since mid-2026, with the National Public Service Portal migrated to the facility in May 2026. Other systems linked to shared national databases are expected to be consolidated there in phases.

The remaining demand, including sector-specific business applications, development and testing environments, backup and disaster recovery, and AI computing capacity, falls within the domain of cloud services, in line with the functional roles defined in the National Digital Architecture Framework.

Cloud adds high-performance computing capacity

The digital transformation strategy sets a target for 100 per cent of shared databases at the national, ministerial, sectoral and local levels to be standardised, connected, access-controlled and ready for AI applications by 2030.

At the same time, 100 per cent of provincial intelligent operation centres (IOCs) are expected to integrate AI for analysis, forecasting and decision support.

AI readiness cannot be achieved simply by developing algorithms or purchasing additional GPUs. A KPMG report found that 48 per cent of public service organisations worldwide expect to deploy AI in production within 12 months, while 43 per cent acknowledge facing major obstacles in scaling AI from pilot projects to real-world applications.

In the public sector, the barriers include data access controls, network infrastructure, information security and the ability to provision computing resources on demand.

To address the infrastructure requirements of large-scale AI, the Government issued Decision No. 1671/QĐ-TTg on August 28, approving the National Artificial Intelligence Strategy through 2030. The strategy formally establishes AI infrastructure as a separate strategic pillar, covering cloud, hybrid/multi-cloud for AI, Edge AI and AI data centres, alongside a National Computing Capacity Coordination Platform.

The strategy sets a target for total AI computing infrastructure capacity of around 560 EFLOPS in BF16 or 1,100 EFLOPS in FP8 by 2030, with the capacity to be invested in by the State, domestic enterprises and foreign enterprises operating in Việt Nam. The total electricity capacity is expected to reach around 500 MW.

The AI Strategy broadens the requirements for cloud infrastructure: beyond general-purpose computing resources, systems must provide high-performance computing (HPC) capabilities and support on-demand GPU integration. This direction is already beginning to be reflected in market investments by major technology groups.

For example, in April 2024, FPT announced plans to invest US$200 million in an AI Factory in collaboration with NVIDIA, equipped with thousands of H100/H200 GPUs. In March 2026, FPT announced plans to deploy NVIDIA HGX B300 systems to support next-generation AI workloads, including reasoning and agentic AI.

A standalone GPU cluster does not, by itself, create AI capacity at an operational scale. AI models can deliver real value only when high-performance computing resources are seamlessly integrated with data, development tools and a secure operating environment.

Cloud sovereignty tailored to data sensitivity

A consistent requirement throughout Decision No. 1266 is technological self-reliance, digital sovereignty, data sovereignty and data security.

In Asia’s public sector, requirements for cloud infrastructure sovereignty are shifting from broad principles to specific regulations based on data sensitivity. The Republic of Korea requires cloud providers serving the public sector to undergo the Cloud Security Assurance Program (CSAP), with requirements for domestic data storage and encryption varying according to system classification.

Malaysia allows public data to be hosted on public clouds, while Restricted data is directed to private or sovereign clouds. For confidential data, policy requires storage within a sovereign cloud zone located in Malaysia.

In Việt Nam, the National Digital Transformation Strategy calls for mobilising private-sector resources and promoting public-private partnerships (PPP) to develop specialised computing infrastructure, while requiring compliance with standards on safety, security, data sovereignty, connectivity, data sharing and shared use.

Some technology companies have begun investing in an approach that combines these layers of capabilities. According to FPT Cloud, its platform has been assessed and approved by the Ministry of Public Security as meeting Level 4 information system security requirements for a cloud platform serving e-government.

The system also meets a set of cybersecurity criteria and complies with TCVN 14423:2026 and Decision No. 8297/QĐ-BCA-A05, covering network, server, application and data security, as well as monitoring and incident response.

FPT’s 'Make in Việt Nam' digital infrastructure also serves as the operational platform for public services of a number of government agencies and ministries.

National digital infrastructure sovereignty should be assessed against measurable and verifiable criteria, ranging from security and access control to monitoring, incident response and business continuity, rather than being determined solely by the physical location of servers or the origin of the service provider.

The gap between 71st place and the global Top 50 in the EGDI cannot be narrowed through the expansion of technical infrastructure alone. The EGDI is built on three pillars: the Telecommunications Infrastructure Index (TII), the Online Services Index (OSI) and the Human Capital Index (HCI).

In effect, the three strategic decisions recently issued by the Government have formally elevated cloud and AI from supporting tools to the foundation of the national digital architecture.

The key challenge for 2026–2030 will be turning the architectural blueprint into tangible improvements in citizens’ experience across individual public services. —VNS

 

E-paper