AI-ready requires four conditions: infrastructure, data, application goals, and people, says Viettel IDC CEO

Investment in AI infrastructure across Southeast Asia has reached tens of billions of USD, while demand for data centers is forecast to grow by around 20% annually in the coming years. The shift is increasing pressure on existing digital infrastructure as data centers move beyond traditional IT workloads to become the backbone for AI applications, which require higher compute density, performance, and scalability.

Cloud platforms reshaped for AI workloads

The transition is forcing cloud platforms to be redesigned to integrate AI capabilities from the infrastructure level. In Vietnam, Viettel IDC is deploying VMware Cloud Foundation 9 (VCF 9), a next-generation cloud platform intended to support AI workloads and serve public cloud, private cloud, and sovereign cloud for enterprises.

Viettel IDC was also named VMware Cloud Service Provider Partner of the Year 2025 by Broadcom in the Asia-Pacific region.

Viettel IDC: AI readiness depends on four prerequisites

In an interview, Viettel IDC CEO Le Ba Tan said the focus should not be on titles, but on how AI is pushing enterprises to change infrastructure approaches—from expanding scale to optimizing compute capacity, and from broad investment to targeted investment in data, energy, and architecture.

Le Ba Tan outlined four prerequisites for being AI-ready:

• Infrastructure: a suitable storage and compute foundation so AI applications can run stably.
• Data: data must be stored and organized systematically so machine learning models can use it effectively.
• Application problems: enterprises must identify where AI can be applied to maximize value, such as boosting productivity, automating processes, improving customer experience, or creating new business models.
• People: leadership and staff must understand the value AI creates, its limits, and how to adjust processes so technology can deliver results.

He added that without all four elements, AI readiness remains limited.

From CPU-centric to GPU-centric data centers

For service providers, AI is changing the data-center landscape by shifting the focus from CPU-centric to GPU-centric compute capability. Moving from CPU to GPU alters power consumption and increases rack power density, which in turn changes electrical distribution and cooling requirements. Many legacy data centers designed for traditional workloads may not be optimal for high-density AI infrastructure.

As a result, data centers are evolving toward an “AI Factory” model—an integrated ecosystem in which hardware, software platforms, AI services, and how value is extracted are closely connected.

Development strategy for 2026–2030

Viettel IDC said it is adjusting its development strategy for 2026–2030 by splitting infrastructure into two layers: a traditional layer to continue upgrading CPU-based services, and a specialized AI layer designed for GPUs, high-density power, and new service platforms. Each data center across three regions will be tailored to local market needs, reflecting the shift from traditional workloads to AI workloads and ensuring reliable power supply.

Le Ba Tan also emphasized that while data centers were previously often viewed in terms of land, AI infrastructure increasingly centers on electricity. A modern data center may not require extensive space, but it must have substantial and stable power capacity. He said green development is becoming necessary because power is a strategic input for digital infrastructure, driving a shift toward clean energy and lower emissions as a long-term requirement.

According to Viettel IDC, the next generation of data centers must meet three simultaneous demands: higher compute density, better energy efficiency, and deeper integration of AI services.

Cost and pricing pressures for users

From the perspective of user organizations, access costs for AI infrastructure remain high. GPU servers cost significantly more than CPU servers, and total cost of ownership includes electricity, cooling, operations, maintenance, and the risk of rapid technology refresh. Because the AI hardware cycle is fast and older equipment loses value quickly, providers face financial risk if demand does not grow fast enough.

As a result, AI infrastructure services are currently selective and primarily suited to organizations with large compute needs, including banks, financial institutions, technology companies, research institutes, and universities. To broaden the market, providers are moving toward flexible pricing—such as by the day or by the hour—rather than fixed monthly or yearly plans, enabling customers to experiment with lower upfront costs and reduced resource waste.

Viettel IDC said that in the early stages, providers may need to invest early and accept suboptimal short-term margins to build the market. Over time, as the customer base expands, the model can become more cost-competitive and sustainable for both providers and customers. It also noted that to bring AI to a broader business community, infrastructure must be strong while services must remain simple and flexible for customers to access.

Domestic strengths and requirements for growth

On domestic versus international cloud providers, Viettel IDC said local players still have substantial opportunity if they address local data residency, industry-specific customization, on-site technical support, integration with existing systems, and transparent pricing. The company argued that Vietnam’s cloud and AI infrastructure market still has room for growth if domestic providers build robust infrastructure and services and understand customer needs.

Regional expansion: “capability” over scale

Viettel IDC said its vision is not limited to the domestic market. It described the Viettel Group approach as expanding internationally after establishing solid capabilities domestically. It said “Go Global” must be grounded in proven operational excellence rather than slogans.

As a first step toward regional expansion, Viettel IDC has launched the Vcloudia brand in Cambodia, leveraging Viettel IDC’s technology base, operational experience, and team, adapted to local enterprise needs.

For overseas expansion, Viettel IDC said its advantage is capability, including:

• platform technology and an international partner ecosystem;
• deep technical expertise;
• real-world operational experience in high-security, high-performance environments;
• adaptability to local market needs.

It said this approach is intended to be more practical for Vietnamese enterprises entering the region than competing solely on scale.

Outlook: skills as the key determinant

Looking ahead 3–5 years, Viettel IDC aims to be a leading digital infrastructure player in Vietnam and to establish a clearer regional position. It said the most critical determinant of whether Vietnam can seize AI opportunities remains people—capability to learn, build, and persevere—adding that policy, capital, and infrastructure are important but difficult to translate into sustainable differentiation without a skilled workforce.

Source: Thanh Niên Việt (Vietnam Youth)