Vietnam has entered the World Bank's upper-middle-income category, marking a significant milestone after nearly four decades of Doi Moi (Renewal) economic reforms.

Yet as the country moves closer to its ambition of becoming a developed economy, sustaining rapid economic growth is becoming increasingly challenging.

To achieve the target outlined by the 14th National Congress - raising GDP per capita to about US$8,500 by 2030 - Vietnam would need to maintain annual economic growth of more than 10% throughout the 2026–2030 period.

With an economy now exceeding US$500 billion, Vietnam is unlikely to reach that goal by relying solely on the growth engines that fueled its success over previous decades.

Instead, the country will need a new source of productivity capable of lifting the entire economy.

It is this challenge that forms the basis of a study led by Professor Tran Tho Dat and a research team from the National Economics University, presented at the National Scientific Conference on Implementing Vietnam's Development Goals to 2030.

Rather than repeating familiar predictions about artificial intelligence, the researchers addressed a different question: How much could AI contribute to Vietnam's economic growth under different policy choices?

Unlike many international reports, the team deliberately avoided pursuing a headline-grabbing figure.

According to Professor Tran Tho Dat, many existing projections estimate substantial economic gains from AI without clearly explaining the underlying models or assumptions.

Instead, the researchers adapted internationally recognized AI economic impact models to Vietnamese economic data in order to construct a series of realistic growth scenarios.

Three possible futures

Economists remain divided over AI's long-term impact on productivity.

While McKinsey has projected that AI could generate enormous value for the global economy, Daron Acemoglu, the 2024 Nobel laureate in Economic Sciences, has argued that AI's productivity gains are likely to be considerably more modest.

Recognizing this uncertainty, the research team chose not to estimate a single definitive outcome. Instead, it modeled AI's potential contribution under Vietnam's specific economic conditions.

The study presents three scenarios.

The baseline scenario assumes AI adoption continues naturally, with little significant policy change.

The accelerated scenario concentrates resources on industries with the highest capacity to adopt AI.

The breakthrough scenario places AI at the center of a coordinated national strategy involving computing infrastructure, data systems, human capital and institutional reform.

The difference between these scenarios lies not in AI technology itself. As AI becomes increasingly accessible worldwide, the decisive factor will be how effectively countries prepare to translate technological advances into higher productivity and, ultimately, faster economic growth.

That, the study argues, is the defining question for Vietnam's next stage of development.

Policy matters as much as technology

During 2026 and 2027, differences among the three scenarios remain relatively small. Investments in AI infrastructure, data, workforce development and institutional reforms all require time before generating measurable economic returns.

After 2027, however, the projected growth paths begin to diverge significantly. Under the baseline scenario, growth reflects the economy's natural pace of AI adoption.

In contrast, the breakthrough scenario gradually pulls ahead as coordinated national investments begin producing cumulative productivity gains.

The model estimates that by 2045, the gap between the baseline and breakthrough scenarios could reach about two percentage points of GDP growth.

The cumulative economic difference could total approximately US$340 billion - nearly equivalent to the current size of Vietnam's economy. The implication is clear: delaying AI-related policies today could come at an enormous long-term economic cost.

For that reason, the researchers describe the period between now and 2028 as a "golden window of opportunity." Every year of delay not only reduces growth in the short term but also diminishes the country's economic potential over the following two decades.

Professor Tran Tho Dat's model therefore measures not only AI's impact but also the impact of public policy.

The same technology, introduced under the same initial conditions, can produce dramatically different economic outcomes depending on policy choices.

Prioritizing sectors with the greatest impact

The researchers do not advocate broad, unfocused investment. Among the 21 industries analyzed, they identify four priority sectors: Information and communications technology (ICT); Finance and banking; Science and technology; Manufacturing.

Although these four sectors account for only about 37% of Vietnam's GDP, they are estimated to generate roughly half of AI's total economic impact.

Given finite public resources, concentrating investment in these industries is expected to create the strongest spillover effects across the broader economy.

This approach also aligns with the objectives of Resolution 57, which identifies science, technology, innovation and digital transformation as key drivers of Vietnam's next phase of development.

However, the study emphasizes that there remains a considerable gap between strategic ambition and measurable economic outcomes.

Bridging that gap will require concrete policies covering AI infrastructure, data governance, human capital, institutional reform and investment.

Ultimately, Professor Tran Tho Dat's research is about more than artificial intelligence.

Its central conclusion is that identical technologies introduced under identical starting conditions can produce very different growth trajectories when accompanied by different policy choices.

If Vietnam wants AI to become its next engine of economic growth, the country's most urgent race is not to acquire the latest technology, but to accelerate institutional reform, strengthen digital infrastructure, expand high-quality data resources and develop the skilled workforce capable of putting AI to productive use.

In the end, the cost of delay will be measured not in technological capability alone, but in the pace of Vietnam's long-term economic growth.

Lan Anh