Asia’s artificial intelligence (AI) investment boom is supporting exports and economic growth across ASEAN+3, but differences in the pace of infrastructure development, productivity gains and financial market adjustments could expose the region to risks if expectations outpace economic fundamentals, according to the ASEAN+3 Macroeconomic Research Office (AMRO).
AMRO said in a commentary on Thursday that the region has emerged as a major beneficiary of rising global demand for semiconductors, advanced electronics and digital infrastructure, with AI-related products providing an important source of export growth amid higher energy costs and persistent global trade uncertainty.
ASEAN+3 accounted for almost half of global exports of AI-enabling goods, including chipmaking inputs, memory and other core computing products, in 2025.
In the first half of 2026, these products contributed about two-thirds of the increase in the region’s merchandise exports, highlighting the growing importance of AI-related demand to its external trade performance.
The research office projects ASEAN+3 economic growth of 4.1 percent in both 2026 and 2027, supported in part by the continued expansion of AI-related investment and demand.
However, AMRO cautioned that the region’s increasing exposure to the global AI investment cycle warrants closer attention to three timing mismatches: between technological investment and power infrastructure development, between investment spending and productivity gains, and between financial market expectations and underlying economic returns.
It said the risks did not necessarily stem from the possibility of the AI boom ending, but rather from expectations moving faster than the economic foundations needed to support the technology’s expansion.
The first mismatch concerns the ability of electricity infrastructure to keep pace with the rapid expansion of AI-related investment.
AMRO noted that while AI capabilities and investment plans could scale up quickly, electricity systems required considerably longer lead times to develop. New power generation capacity and transmission networks involve years of planning, regulatory approvals, financing and construction.
This creates risks in both directions. A shortage of electricity infrastructure could result in higher operating costs or constrain the operations of data centres and other AI-related facilities. Conversely, building infrastructure based on overly optimistic demand projections could leave economies with expensive assets that are underutilised.
The issue is particularly relevant to Southeast Asia, where rising electricity demand from AI-related activities is adding to consumption growth driven by manufacturing expansion and urbanisation.
In Malaysia, applications for data center power supply had exceeded 11 gigawatts as at December 2024, equivalent to about 40 percent of Peninsular Malaysia’s generating capacity, illustrating the scale of potential electricity requirements from the sector.
AMRO said policymakers would therefore need to ensure that power infrastructure development was aligned with realistic demand projections, while maintaining reliable electricity supply and avoiding excessive investment in capacity that might not be fully utilised.
The second mismatch is between the immediate economic benefits generated by AI investment and the longer period needed for productivity improvements to materialise.
Investment in advanced chip manufacturing, data centers and computing equipment generates orders, revenues and economic activity relatively quickly. However, the productivity gains associated with AI adoption depend on broader changes in business processes, workforce skills and organisational structures, as well as the ability of companies across different industries to integrate the technology into their operations.
These adjustments are neither automatic nor immediate, meaning economies could experience a strong investment-driven expansion before seeing comparable improvements in overall productivity.
AMRO’s analysis of the International Monetary Fund’s AI Preparedness Index highlights significant differences in the region’s ability to capture these benefits.
Singapore, Japan, South Korea and Hong Kong scored above 70 on the index’s 0-to-100 scale, placing them in the same group as the United States. In contrast, six of AMRO’s 14 member economies, including the Philippines and Vietnam, scored below 50.
The differences suggest that economies with lower AI readiness could take longer to translate investment into widespread productivity improvements.
AMRO said this gap did not necessarily mean that AI-related investment was misguided. However, it created uncertainty over when the expected economic returns would materialise and how widely the benefits would be distributed across industries and businesses.
The third mismatch involves financial markets, which tend to respond to expectations of future earnings much faster than infrastructure can be built or productivity gains can be realised.
Equity valuations and financing decisions are influenced by expectations of future corporate earnings. A revision to those expectations could therefore trigger sudden changes in financial conditions, even if the underlying technological transformation remained intact.
Recent volatility in global AI and semiconductor stocks illustrates how quickly market valuations could be repriced.
For ASEAN+3, the potential impact extends beyond individual technology companies. Given the region’s integration into global production networks, changes in AI investment plans could affect semiconductor exports, equipment orders, corporate earnings, equity prices and cross-border capital flows simultaneously.
AMRO warned that these developments could reinforce one another, with weaker export demand affecting corporate revenues, while falling asset valuations and tighter financing conditions could discourage further investment.
Economies with substantial exposure to the same global technology cycle could consequently face adjustments at around the same time.
To illustrate the potential economic impact, AMRO’s scenario analysis examined the implications of a slowdown in US technology investment.
If US technology investment were to return to its more moderate 2024 growth pace, instead of maintaining its recent momentum, ASEAN+3 economic growth in 2027 could be 1.5 percentage points below the baseline projection.
AMRO stressed that this was a scenario analysis rather than a forecast, and did not imply that AI investment was inherently unsustainable.
It acknowledged that AI could deliver substantial and lasting productivity improvements, but cautioned against assuming that the transition from technological innovation to economic returns would be smooth or that financial markets would adjust gradually.
A technological transformation could still be genuine even if some expectations surrounding its economic potential proved premature, it said.
Against this backdrop, AMRO identified several priorities for policymakers to manage the risks while allowing the region to benefit from the AI boom.
First, governments should ensure that essential infrastructure expands in line with realistic demand projections. Public investment and regulatory frameworks should support reliable electricity supply while ensuring that private investors continue to bear appropriate commercial risks.
Second, policymakers should focus on translating AI investment into broader productivity gains by strengthening workforce skills, encouraging technology adoption among smaller businesses and creating conditions for innovation to spread beyond leading technology exporters.
Third, financial supervisors should monitor vulnerabilities beyond direct exposure to technology companies.
Potential risks could also emerge through suppliers, infrastructure financing, equity markets and non-bank financial intermediaries, including insurers and private credit funds.
AMRO said financial surveillance should take into account the possibility that a common shock could affect several of these channels simultaneously, potentially amplifying the impact on regional economies.
It also highlighted the importance of regional cooperation in identifying shared vulnerabilities and facilitating information exchange, although individual economies would still need to assess their own exposure and maintain sufficient policy flexibility to respond to unexpected developments.
AMRO said Asia had good reason to welcome the opportunities created by AI, given its manufacturing capabilities, technological expertise and established production networks, which positioned the region to benefit from the global transformation.
However, sustaining this advantage would require more than attracting investment and expanding computing capacity.
Ultimately, the success of Asia’s AI boom would depend not only on the pace of investment growth, but also on how effectively the region translated technological advances into lasting economic value while managing the associated infrastructure, productivity and financial risks.
Thailand’s AI investment boom offers chance to revive long-term growth, says AMRO