“AI is shifting from being a standalone technology sector to becoming a form of general-purpose infrastructure. If we fail to view capital for the AI economy as a multi-layered ecosystem—PPP for AI infrastructure, venture capital for startups, and credit/incentives for applications—then AI in Vietnam will remain at the stage of small-scale pilots,” said Dr. Vu Hoang Linh from the University of Economics, Vietnam National University, Hanoi, in an interview with The Saigon Times.

An ambitious scenario and the funding question The Saigon Times: According to the Vietnam AI Economy Report 2025 released in July 2025, by 2040 artificial intelligence (AI) could contribute up to US$130 billion to Vietnam’s GDP, equivalent to about 25% of the economy’s current size. How do you view this projection? Are we being overly optimistic? Dr. Vu Hoang Linh: The figure of US$130 billion by 2040 is indeed ambitious. If measured against today’s GDP (forecast at around US$491 billion in 2025), that’s 25%. But if measured against Vietnam’s projected GDP in 2040 (about US$1,200–1,300 billion, assuming 6–6.5% annual growth), AI’s contribution would be around 10–11%. That’s still very large—roughly triple Vietnam’s current total textile export turnover. However, to reach this level, Vietnam must focus on four critical areas:

1. Computing and energy infrastructure: AI’s GDP contribution mainly comes from large-scale applications. Without data centers, GPUs, and stable electricity, deployment capacity will remain constrained. Singapore has nearly 1,000MW in data center capacity, while Vietnam only has around 200MW. Without infrastructure investment, the US$130 billion will remain on paper.
2. Data and legal framework: AI learns from data. If public and private datasets remain fragmented, lack open standards, and face barriers like cross-border data transfer restrictions, businesses will lack the “raw material” to create value. GDP contribution will struggle to spread across the economy.
3. Human resources: Vietnam has many basic software engineers but lacks applied AI engineers. If this gap isn’t bridged, AI-driven productivity gains will concentrate in a few large companies—insufficient to generate US$130 billion.
4. Market-driving demand: The US$130 billion projection assumes AI is widely applied in public administration, healthcare, education, and manufacturing. If the Government does not launch large-scale, leading projects, private enterprises will lack the incentive to invest. Without that, AI cannot spread at the macroeconomic level.

In short, the US$130 billion target is plausible in terms of potential—but only if Vietnam simultaneously addresses these four bottlenecks. Otherwise, the forecast leans more toward optimism than reality. You have said that AI is not just a distinct technology sector, but a foundational layer with the ability to reshape all fields. Can we then understand AI as an infrastructural element of the economy? In this case, how should capital for the AI economy be viewed? - AI is shifting from being a standalone technology sector to general-purpose infrastructure, similar to electricity or the Internet. This means that when we discuss capital for the AI economy, we cannot only talk about venture funds for startups. Instead, we must organize it into three distinct capital layers. First, foundational infrastructure: Vietnam has seen several large-scale projects, such as FPT’s US$200 million AI Factory; the initial phase of CMC’s Hyperscale Data Center worth US$250 million, with potential expansion to about US$1 billion; and VNPT’s IDC Hoa Lac, the largest Tier III-certified data center in Vietnam. However, the overall scale is still modest compared with Singapore, whose data center capacity exceeds 1,000MW—five times Vietnam’s. This shows Vietnam must accelerate infrastructure investment if AI is to truly become foundational. Second, startups and R&D: In 2024, total innovation investment into Vietnam reached US$2.3 billion across 141 deals (NIC, 2025). The share of funding for AI R&D and deeptech rose sharply, as shown by the multiple increase in capital flowing into AI startups compared to the previous year. This is a positive signal, indicating that international investors see Vietnam as a promising market for applied AI. Third, public-private partnership (PPP) applications: Data shows that public-sector IT spending typically accounts for 1.5–2% of recurrent expenditure, while large Vietnamese enterprises spend an average of 1.8–2.2% of revenue on IT (Vietdata Report 2024). A growing portion of this IT budget is allocated to AI, especially in finance, retail, and logistics. If the Government introduces tax incentives and expands public procurement under an “AI ordering” model, application capital will spread rapidly. In summary, if we fail to view capital for the AI economy as a multi-layered ecosystem—PPP investment for infrastructure, venture capital for startups, and credit/incentives for applications—then AI in Vietnam will remain stuck at the stage of small-scale pilots. Conversely, if capital is structured properly, AI can become a core infrastructural layer of the economy. Vietnam’s advantages Looking at each component—capital for AI infrastructure investment, capital for AI startup innovation, and capital for AI application in various sectors—which area does Vietnam have an advantage in? And in your view, where should we place priority so that it both aligns with Vietnam’s economic conditions and potential, and also yields the greatest impact? - In reality, Vietnam’s clearest advantage today lies in AI application. According to a late-2024 survey by NIC, about 19% of small and medium enterprises (SMEs) had tried integrating an AI product, but fewer than 4% maintained long-term use. Meanwhile, a CPA Australia report shows that 44% of small businesses in Vietnam considered AI their top tech investment in 2024, and over 47% were already using AI for core operations. This indicates that demand for application has become strong, particularly in customer service, logistics, and manufacturing. By contrast, in computing infrastructure, we are still weak. The Ministry of Science and Technology reported that by 2025, Vietnam will only have 221MW of operating data center capacity. This gap explains why compute costs remain a major bottleneck for domestic AI firms. As for scientific R&D, Vietnam’s contribution to the AI field remains modest. According to Scopus, AI papers from Vietnam account for less than 1% globally, reflecting limitations in foundational research and academic workforce. In this context, Vietnam’s reasonable strategy should be to

1. Focus on large-scale applications in public administration, education, and healthcare—the three areas recommended by NIC—to create markets and generate quality data.
2. Make targeted investments in green compute (data centers powered by renewable energy) to ensure minimum self-reliance and reduce dependence on international infrastructure.
3. Prioritize training 10,000–20,000 applied AI engineers over the next five years, concentrating on practical applications and integrating AI into production and service processes rather than solely focusing on theoretical research.

With these three pillars—“application pull, infrastructure push, and human resources as foundation”—Vietnam can leverage its current strengths, match its economic conditions, and create the widest spillover effects. In your opinion, what is the biggest bottleneck in mobilizing capital for the AI economy, and how can we resolve it? - The biggest obstacle is that intangible assets are difficult for banks to accept as collateral. Data, algorithms, and intellectual property rights are not easily collateralized, making it hard for AI companies to access long-term capital. At the same time, computing infrastructure costs are extremely high, while Vietnam’s electricity prices are expected to keep rising in the near future. To address this, we need:

1. A dedicated legal framework allowing data and AI software to be recognized as valuable assets.
2. Blended finance: a model combining Official Development Assistance/Development Finance Institution (development banks, international funds) with commercial banks, with government guarantees for AI infrastructure. For example, a US$300 million AI data center project could be structured as: 40% concessional ODA loans, 40% domestic commercial loans, and 20% corporate equity. If the State guarantees 30% of the commercial loan portion, risk is reduced and banks will lend more readily.
3. Tax deduction policies for AI adoption costs by SMEs (which currently account for 98% of all businesses).

If the capital–intangible asset challenge is not resolved, it will be very difficult to scale the AI economy across the entire economy. Concrete action needed Finally, you have previously suggested that Vietnam has the capacity to turn AI into a true driver of development. At this point, how would you answer that question? Could you give more specifics? - I believe Vietnam absolutely has the potential to turn AI into a genuine driver of development — provided that policies are translated into concrete action. On the legal side, the Digital Technology Industry Law (DTI Law) was just passed by the National Assembly in June 2025, opening a legal framework for AI and semiconductor investment. Almost simultaneously, the Personal Data Protection Law (PDPL) was also passed, effective from early 2026, creating the necessary legal foundation for data and AI. On infrastructure, major corporations have already taken concrete steps: FPT is rolling out its AI Factory, while CMC is partnering with Samsung C&T to build a hyperscale data center in Ho Chi Minh City. On market demand, there are also encouraging signs: ministries have begun deploying large-scale AI solutions in public administration, education, healthcare, and urban transport. For instance, Ho Chi Minh City launched a citizen-assisting robot and integrated ChatGPT at its administrative center in July 2025, reducing processing times and improving public service experiences. The Ministry of Health has set a target of training 15,000 healthcare workers in AI between mid-2025 and 2026, with applications in clinical practice, management, and research. In transportation, AI and satellite data have been widely deployed to analyze school traffic accident risks since 2021–2024, and expansion nationwide is underway. On human resources, many universities have launched AI programs—for example, four economics-oriented universities (National Economics University, University of Economics Ho Chi Minh City, Ho Chi Minh City Banking University, and University of Economics and Law). The Posts and Telecommunications Institute of Technology has opened the first dedicated AI faculty. Hanoi University of Science and Technology is partnering with NVIDIA and NIC for advanced AI training. If these four pillars—legal framework, infrastructure, market demand, and human resources—are implemented consistently and effectively, then AI will not remain just a slogan, but will genuinely become a new driver of productivity, with the potential to make significant contributions to GDP over the next 15–20 years.