Artificial intelligence (AI) investment in Southeast Asia is entering an unprecedented boom, projected to exceed US$110 billion by 2028. From sprawling data centres and next-generation generative AI platforms to cloud infrastructure and automation start-ups, capital is flowing rapidly into the region’s digital future. Yet, amid the optimism, economists and policymakers are grappling with a central question: will this wave of investment yield the productivity revolution that justifies its cost—or is it laying the groundwork for another speculative bubble?
The surge in AI-related investment across Southeast Asia is staggering. Major economies such as Singapore, Indonesia, Malaysia, and Thailand have already attracted US$55.2 billion in AI commitments, distributed as follows: Singapore (US$10 billion), Indonesia (US$7.4 billion), Malaysia (US$23.2 billion), and Thailand (US$8.4 billion). Analysts forecast this figure will double within four years, expanding at a compound annual growth rate of 25 per cent, ultimately surpassing US$110 billion by 2028.
This regional trend mirrors the global AI investment explosion, led by the United States, where five major technology firms—Microsoft, Google, Amazon, Meta, and Apple—plan to spend more than US$325 billion on AI initiatives in 2025 alone. With over 400 million digital consumers generating US$263 billion in digital economy revenue in 2024, Southeast Asia has become an attractive investment destination for both U.S. and Chinese tech giants seeking access to young, connected markets.
Beyond demographics, geography is also a factor. The region’s location—bridging China, India, and Australia—positions it as a strategic hub for digital and data networks, making it an ideal testing ground for cross-border AI innovation and integration.
Despite the flood of capital, the economic rationale behind these investments remains uncertain. AI promises to reshape industries, enhance efficiency, and open new markets, but these benefits hinge on one critical condition—a measurable improvement in labour productivity.
In other words, for Southeast Asia’s US$110 billion bet on AI to make economic sense, the technology must produce gains in productivity large enough to justify its cost. But how large must these gains be—and are they realistically achievable?
Economists use computable general equilibrium (CGE) models to estimate the productivity growth required to match expected returns on such massive investments. These models provide a benchmark for evaluating whether current projections align with past performance and future potential.
In advanced economies, particularly the United States, economists are divided on how much AI will actually improve productivity.
Optimists point to significant potential gains. Research from the Brookings Institution, Federal Reserve Bank of St. Louis, and Goldman Sachs estimates that generative AI adoption could lift U.S. annual labour productivity growth by 1.1 to 1.8 per cent. Likewise, economists Philippe Aghion and Simon Bunel predict a 0.68 to 1.3 per cent annual rise in total factor productivity (TFP)—a broad measure of efficiency in combining labour and capital.
However, sceptics remain unconvinced. Prominent MIT economist Daron Acemoglu estimates that AI’s total contribution to TFP will be just 0.71 per cent over ten years, or roughly 0.07 per cent per year—barely a dent compared to the productivity growth seen during the industrial or digital revolutions. McKinsey & Company offers a middle ground, projecting annual productivity increases of 0.1 to 0.6 per cent from AI adoption.
This divergence underscores a deeper uncertainty: while AI can transform tasks and processes, translating those transformations into aggregate productivity gains—especially in developing economies—may prove elusive.
Applying CGE models to Southeast Asia reveals a sobering picture.
For current AI investments to yield economically sound returns, labour productivity must increase by an additional:
- 0.79 per cent annually in Indonesia
- 2.5 per cent annually in Malaysia
- 6.52 per cent annually in Thailand
To contextualise, the average annual labour productivity growth from 2014 to 2024 was roughly 2 per cent across Indonesia, Malaysia, and Thailand. That means AI would need to deliver productivity boosts 38 per cent higher in Indonesia, 130 per cent higher in Malaysia, and over 320 per cent higher in Thailand than historical trends.
Put plainly, unless AI unleashes transformational gains, much of this investment could struggle to meet expectations.
If the AI pessimists are right, only Indonesia’s relatively modest targets appear achievable. For Malaysia and Thailand, by contrast, the success of these investments depends heavily on AI optimists being correct about large, rapid productivity improvements.
Even if AI adoption does not translate directly into measurable productivity growth, it may still generate transformative effects across society and industry. Economists recall the “Solow paradox” of the 1980s, when Nobel laureate Robert Solow observed that “you can see the computer age everywhere but in the productivity statistics.”
Similarly, AI’s benefits—improved decision-making, faster innovation, enhanced customer service—may not immediately appear in traditional metrics. Productivity data often lags technological diffusion, and in economies where digital infrastructure and workforce skills are still developing, that lag can be even longer.
Regardless of whether AI produces immediate statistical gains, governments have a pivotal role in ensuring that these investments translate into sustainable growth. Economic history suggests that technology alone does not drive productivity—it must be accompanied by institutional and structural readiness.
- Adoption, Not Just Investment:
Many Southeast Asian firms, especially small and medium enterprises (SMEs), lack the digital literacy or resources to integrate AI tools effectively. Governments must incentivise broad-based adoption through digital training, tax credits, and open-access innovation platforms. - Human–AI Complementarity:
Productivity gains depend on how well AI complements human skills rather than replacing them. Policies promoting education reform, STEM training, and lifelong learning are crucial to aligning human capital with AI-driven economic structures. - Openness to Trade and Foreign Investment:
AI development thrives in open, interconnected markets. Restrictions on data flows, foreign direct investment, or technology partnerships could stifle innovation. Maintaining openness to both U.S. and Chinese tech ecosystems will be essential for Southeast Asia’s success. - Creative Destruction and Economic Dynamism:
Allowing unproductive firms to exit and new, innovative firms to rise—a process known as creative destruction—ensures that capital and talent flow to where they are most efficient. Overregulation or protectionism could slow this natural reallocation process.
For Southeast Asian governments, the challenge is not to decide whether AI will succeed or fail—but to ensure that whatever its eventual impact, the region is ready to harness it productively. Policymakers should focus less on speculative forecasts of AI’s power and more on creating the conditions for adaptability.
This includes building robust digital infrastructure, strengthening intellectual property regimes, promoting cross-border data governance, and supporting regional AI research collaborations under frameworks such as ASEAN’s Digital Economy Framework Agreement (DEFA).
Ultimately, AI investment in Southeast Asia is both a risk and an opportunity. If current projections prove overly optimistic, the region could face an “AI bubble” reminiscent of past technology booms that failed to deliver expected returns. But if governments and industries align their strategies toward inclusive adoption, skill development, and innovation ecosystems, these investments could underpin a new era of productivity-led growth.
As one regional economist recently observed, “Southeast Asia shouldn’t be an AI optimist or AI pessimist—it should be an AI pragmatist.”
That pragmatism means recognising that AI’s payoff is not guaranteed by capital alone. Its success will depend on people, policies, and partnerships that convert technological promise into measurable prosperity.
In that sense, the region’s US$110 billion AI gamble is not just a question of economics—it is a test of strategic foresight.
