According to Le Hong Viet, CEO of FPT Smart Cloud, the gap in artificial intelligence (AI) investment between Vietnam and global powers like the United States and China represents a “tremendous challenge.”

nguyen hoang hung viettel

Nguyen Hoang Hung, CTO of Viettel AI, speaking at the AI4VN event on September 26.

In the global race for AI dominance, computing infrastructure is considered a foundational pillar. At the AI4VN 2025 event, experts acknowledged that while Vietnam faces significant hurdles, it also has major opportunities to break through.

Nguyen Hoang Hung, Chief Technology Officer of Viettel AI, shared that top global tech giants like Meta, Microsoft, and OpenAI have invested tens to hundreds of billions of US dollars to acquire millions of GPUs.

Training Meta’s Llama 3 model, for instance, required 30.84 million GPU hours - a workload that would take up to 55 years to complete using small-scale domestic server clusters. This figure highlights Vietnam’s massive shortfall in AI infrastructure compared to world leaders.

Meanwhile, Le Hong Viet pointed out that Vietnam’s data center market is still relatively small and unable to compete with other regional countries. The country’s AI investment remains modest, trailing the US and China by 56 times - and even falling behind Singapore.

Currently, the US leads the global AI landscape with 40 developed models and private investment reaching USD 471 billion between 2013 and 2024. China follows closely with 15 high-quality AI models and the world’s largest number of AI-related patents, narrowing the quality gap.

Beyond funding disparities, Vietnam also faces challenges such as a shortage of high-quality AI talent, inadequate R&D spending, and an evolving legal framework for artificial intelligence.

Bright spots in Vietnam’s AI landscape

le hong viet fpt
Le Hong Viet, CEO of FPT Smart Cloud, presents his strategy for Vietnam’s AI breakthrough. Photos: Giang Huy

Despite the challenges, Vietnam shows notable signs of progress. A report from the Worldwide Independent Network of Market Research (WIN) ranks Vietnam 6th out of 40 countries in terms of readiness for the AI era. The domestic AI ecosystem is rapidly gaining traction, attracting USD 80 million in investment in 2024, supported by a workforce of 500,000 tech professionals and a high adoption rate - 42% of the population and 65% of small and medium enterprises (SMEs) have used AI.

The government has set a target for Vietnam to become one of the top three countries in Southeast Asia and among the top 50 worldwide in AI research by 2030. Local tech enterprises are stepping up investments, affirming the perspective that "AI infrastructure sovereignty is not just a corporate responsibility."

Viettel, for example, currently operates 15 data centers and has invested in cutting-edge GPU systems such as Nvidia DGX SuperPOD. The company is also developing a “Make in Vietnam” ecosystem of AI products serving tens of millions of users.

Vietnamese businesses are also proactively adopting AI Agents to transform operations. According to FPT Smart Cloud, more than 1,500 AI Agents have been deployed, automating 46% of customer service center workloads, boosting telesales revenue by 20%, and processing over 400 million documents annually with accuracy exceeding 95%.

Le Hong Viet emphasized that Vietnam needs a strategy to “build AI with sovereignty,” aiming not only to catch up but also to lead within ASEAN. His proposed 2025–2030 roadmap includes three phases: foundation building, expansion, and regional leadership, focusing on four key pillars - human capital, digital infrastructure, product innovation, and ecosystem development.

Experts agreed that while infrastructure limitations are a major challenge, they also serve as motivation for Vietnam to find its own path. With competitive costs, a large pool of young tech talent, and strong government backing, a well-executed investment strategy could help Vietnam narrow the gap with global powers and emerge as a rising AI hub in the region.

Du Lam