Asia’s AI Decade - The Super Intelligence Overlay for Asia

Asia’s AI Decade

I believe Asia is entering a period in which the traditional measures of economic advantage—population size, low-cost labor, education levels and even GDP—will become substantially less useful on their own. Artificial intelligence is changing the production function of an economy. Over the next decade, the countries that benefit most will not necessarily be those with the largest populations or the highest average educational attainment. They will be those that combine human capability, creativity, capital, reliable and affordable energy, infrastructure, industrial capacity and the ability to deploy AI at scale.

The central question I would ask as an investor, entrepreneur or policymaker is therefore not simply, “How many people does this country have?” It is: “How much productive capability can each person generate when augmented by AI?”

That distinction could profoundly reshape the Asian economic hierarchy.

AI will not make human capital irrelevant—it will change what kind of human capital matters

I initially viewed education as one of the principal advantages in the AI race. I now think that is incomplete.

AI can increasingly provide many of the capabilities that previously required years of specialized training: coding, research, translation, analysis, design, accounting, legal research and technical assistance. This does not eliminate the value of educated workers, but it potentially reduces the economic premium attached to routine intellectual execution.

What becomes more scarce is the ability to determine what should be built, what problem is worth solving, how to combine disparate technologies, and how to turn an idea into a commercially viable product or company.

In other words, I would distinguish between structured intelligence and creative intelligence.

A highly educated population that is exceptionally good at following established processes, optimizing systems and operating complex institutions has an enormous advantage today. But an AI-augmented economy may increasingly reward experimentation, entrepreneurship, invention, unconventional thinking and commercialization.

This distinction changes how I evaluate several Asian economies.

Singapore illustrates the distinction particularly well

Singapore is an extraordinary example of structured human capital, institutional competence and economic execution. It has world-class education, infrastructure, finance, logistics, digital connectivity and policy execution. The IMF currently places Singapore at the top of its AI-preparedness framework, while WIPO ranks it fifth globally in innovation.

But I would be careful about equating Singapore's educational attainment with exceptional endogenous creativity.

WIPO's 2025 data are revealing: Singapore ranks first globally in innovation inputs but ninth in innovation outputs, with creative outputs ranking 15th.

That does not mean Singapore lacks innovation. It means its comparative advantage has historically been particularly strong in organizing capital, talent, infrastructure and institutions to produce economic results.

That is different from producing the world's largest number of new consumer technologies, cultural movements, global brands or disruptive entrepreneurial ecosystems.

Singapore may therefore remain one of Asia's most efficient AI adopters without necessarily being the place where the greatest number of entirely new ideas originate.

South Korea demonstrates another model

South Korea is a useful counterexample because it has demonstrated not only educational and technical capability but also significant technological and cultural creation.

Its industrial ecosystem has produced globally competitive companies in semiconductors, electronics, automobiles, batteries, telecommunications and consumer technology. It has also produced globally distributed cultural products in music, film, television, gaming, cosmetics and fashion.

WIPO ranks Korea fourth globally in innovation and first in its human-capital-and-research pillar. Korea also ranks among the global leaders in business R&D, researchers employed by business and patent activity.

That matters because AI is unlikely to reward only the ability to operate an existing system. Countries capable of creating new systems, products and industries may have an additional advantage.

Demographics must therefore be evaluated differently

I also believe we should stop treating population growth as an automatic economic advantage.

A country with 100 million relatively low-productivity workers does not necessarily have a demographic advantage over a country with 30 million highly productive workers.

A more useful formulation is:

Economic productive capacity ≈ workers × human capital × AI capability × capital × infrastructure × energy × organizational capability.

The Philippines and Indonesia, for example, have enormous demographic assets. But the economic value of those populations depends on whether education, infrastructure, electricity, capital and technology can convert population into productivity.

Conversely, Japan, South Korea and Taiwan face severe demographic contraction, but their existing human capital, capital stock, industrial ecosystems and technological capabilities allow each worker to produce substantially more economic value.

The IMF's recent work on AI and economic divergence in Asia reinforces this point: countries with stronger infrastructure, capital intensity and human capital are positioned to adopt AI earlier, while demographic and capital constraints can delay adoption elsewhere.

The demographic question I therefore care about is not simply fertility.

It is:

How productive will the next generation of workers become when combined with AI?

This is where Vietnam becomes particularly interesting

Vietnam may have one of the most compelling combinations in Asia.

It has a large working-age population, a rapidly expanding manufacturing base, substantial foreign investment, improving infrastructure and an educational foundation that is stronger than its income level might suggest.

The country has increasingly positioned itself within global electronics and manufacturing supply chains. That gives Vietnam something that pure demographic statistics cannot capture: a mechanism for turning human capital into industrial capability.

Its opportunity is to move from:

low-cost manufacturing → increasingly sophisticated manufacturing → domestic engineering capability → technology companies → AI-enabled productivity.

If that transition continues, Vietnam does not need to become Singapore. It needs to become increasingly productive.

That is a very different objective.

Malaysia may have an equally interesting AI infrastructure opportunity

Malaysia has a smaller population than Indonesia or the Philippines, but I do not view that as a disadvantage.

Its combination of relatively strong education, electronics and semiconductor capabilities, industrial infrastructure, energy resources, proximity to Singapore and rapidly expanding data-center industry gives it an unusual position.

Malaysia is increasingly becoming an important regional destination for digital infrastructure. The World Bank has documented a major expansion of data-center investment and electricity commitments in the country.

This creates a potential virtuous cycle:

energy + data centers + semiconductor/electronics expertise + foreign investment + skilled workers → higher-value digital and industrial activity.

Malaysia's opportunity is not to win through population scale. It is to maximize productivity per worker and productivity per unit of energy and capital.

Indonesia represents the opposite proposition

Indonesia has what Malaysia lacks: enormous demographic and domestic-market scale.

With hundreds of millions of people, abundant natural resources and a growing digital economy, Indonesia has tremendous potential.

But its challenge is converting population into productive human capital.

If AI can substantially raise the productivity of workers who currently lack access to high-level expertise, Indonesia could experience enormous gains. AI can potentially function as a form of distributed expertise—putting sophisticated analytical, educational and technical capabilities into the hands of millions of people.

That makes Indonesia one of the most important AI diffusion stories in Asia.

Its opportunity is enormous, but its outcome will depend heavily on education, connectivity, electricity, infrastructure and capital formation.

The Philippines presents a different problem

The Philippines possesses one of Asia's strongest demographic profiles, a large English-speaking workforce and an established services and BPO industry.

But precisely because so much of its economic model depends on human labor performing information-intensive services, AI creates an unusual structural risk.

AI does not necessarily destroy the BPO industry. It could instead allow a company to generate the same revenue with substantially fewer workers.

That distinction matters enormously.

A growing BPO industry can therefore coexist with a declining requirement for BPO labor.

The Philippines also faces comparatively high electricity costs and infrastructure constraints. The World Bank has identified electricity affordability and grid investment as important competitiveness issues, while also identifying substantial potential from renewable-energy development.

The Philippines therefore has an extraordinary demographic asset—but it must convert that asset into AI-complementary human capital and lower-cost productive infrastructure.

Energy may become one of the most underappreciated determinants of AI competitiveness

The AI economy is not purely digital.

It is extraordinarily physical.

Data centers require electricity. Semiconductor fabrication requires electricity. Advanced manufacturing requires electricity. Cooling, networking, robotics and industrial automation all require reliable power.

As AI adoption accelerates, I expect reliable, scalable and affordable electricity to become a more important determinant of national competitiveness.

This creates an unusual opportunity for countries such as Vietnam, Malaysia and Indonesia, provided they can expand generation and grids efficiently.

It also creates a structural challenge for economies where electricity is expensive or unreliable.

The countries that win the AI infrastructure race will not simply have the best models. They will have the ability to power the models and the factories that use them.

My emerging view of the Asian hierarchy

I would therefore divide the region into several different categories rather than simply ranking countries by GDP or population.

Singapore, Taiwan, South Korea and Japan possess extraordinary existing technological, capital and human-capital advantages. Their principal challenge is demographic aging and, in some cases, energy cost.

Malaysia and Vietnam have particularly interesting combinations of demographics, manufacturing, education, infrastructure, foreign investment and AI opportunity. They may be among the region's most interesting incremental beneficiaries.

Indonesia has enormous upside because of scale, resources and demographics, but its outcome depends heavily on raising productivity and human capital.

Thailand has significant industrial capabilities and a substantial existing economic base, but faces demographic aging, productivity challenges and the need to transition toward higher-value activities.

The Philippines has strong demographics and service capabilities but faces greater exposure to AI-driven labor substitution and relatively high energy costs.

Cambodia, Laos and Myanmar face much larger gaps in education, infrastructure, capital and institutional capacity. AI can still be transformative in these economies, but diffusion constraints are considerably greater.

These are not predictions of GDP rankings. They are observations about the structural conditions that determine how effectively each economy can convert AI into productivity.

The central investment thesis

My central thesis is therefore simple:

The winners of Asia's AI decade will not necessarily be the countries with the most people, the cheapest labor or even the highest educational attainment. They will be the countries that combine educated and creative people with capital, affordable energy, infrastructure, industrial capability and the institutional ability to deploy AI at scale.

I would go one step further.

AI may ultimately make creativity more economically valuable relative to routine intelligence.

Education remains essential. But education that merely produces competent executors may become less differentiated as AI becomes more capable.

The most valuable human being in an AI economy may increasingly be the person who can say:

“Here is a problem nobody else has solved. Here is a product nobody has built. Here is a market nobody has recognized. Now let's use AI to build it.”

That is why I would not evaluate Asia's next decade solely through fertility rates, population growth, university rankings, GDP per capita or AI-readiness indexes.

I would evaluate each country according to a broader equation:

AI-era national productivity = human capability × creativity × capital × energy × infrastructure × industrial capacity × institutional execution.

The countries that can maximize all of those variables simultaneously may look very different from today's economic hierarchy.

And that, in my view, is the real Asian investment story of the next decade.