SE Asia lags in AI boom despite manufacturing presence

Edward Lee, Standard Chartered's chief economist and head of FX for ASEAN and South Asia, told Fortune that despite heavy foreign direct investment flows, Southeast Asia has failed to climb the value chain in the AI era. The region contributes just 6% of global intermediate manufacturing, versus 15% for China—a gap that reflects deeper structural disadvantage.

The AI hardware boom is concentrating wealth in North Asia. TSMC reported 707 billion New Taiwan dollars ($22 billion（約3.5兆円）) in quarterly profit on July 16, a 77% year-on-year jump that significantly beat analyst expectations. Samsung Electronics and SK Hynix are also capturing outsized gains from surging demand for AI processors and their components. Southeast Asia does participate in the AI supply chain: Singapore produces semiconductors, Malaysia handles chip assembly, packaging, and testing, and both Malaysia and Thailand assemble AI servers. However, Lee notes, "the AI sector in Southeast Asia is significantly smaller than in North Asia, home to the leading semiconductor companies, AI server assemblers, and device component manufacturers."

Lee argues that Southeast Asian governments and companies must invest in AI-related research and development to prevent irrelevance. Singapore has led by example: the Singapore government committed over 1 billion Singapore dollars ($776 million（約1200億円）) to fund fundamental AI and applied research. However, Lee cautioned that "R&D is costly." He also urged caution about generalizing AI's regional impact. Singapore, Malaysia, and Vietnam are likely to benefit from increased demand for electronics manufacturing, he predicted. But the Philippines faces a different calculus. Its business process outsourcing industry contributes 8% of GDP and employed 1.89 million people in the most recent year, generating $40.3 billion（約6.4兆円） in revenue. On July 14, the IT & Business Process Association of the Philippines downgraded its best-case 2028 scenarios: revenue fell from a $59 billion（約9.4兆円） forecast to $50.5 billion（約8.1兆円）, and projected headcount dropped from 2.5 million jobs to 2.14 million. The downgrades reflect concern that AI will automate roles currently filled by Filipino workers.

When asked whether Southeast Asia could replicate the success of established tech hubs, Lee was skeptical. "Every country wants to have the next Silicon Valley or be the next AI mega-producer, but getting there isn't just a matter of snapping one's fingers or throwing money towards the industry—that may not work," he said. "Unfortunately, I don't have a silver bullet solution."

Southeast Asia's struggle to capture value in the AI boom reflects a broader structural challenge: while the region has manufacturing capabilities in semiconductors, chip assembly, and server assembly, it lacks the integrated supply chain dominance and technological depth of North Asia. Taiwan, South Korea, and Japan have consolidated control over high-value segments—TSMC alone reported 707 billion New Taiwan dollars ($22 billion（約3.5兆円）) in quarterly profit in July, a 77% year-on-year jump—while Southeast Asia's contributions remain peripheral. Edward Lee's observation that the region "has not really moved up the value chain" despite decades of foreign direct investment highlights the risk of remaining locked in lower-margin assembly and testing roles.

The Philippines case illustrates how AI disruption compounds existing vulnerabilities. The business process outsourcing sector, which employs 1.89 million people and generated $40.3 billion（約6.4兆円） in revenue last year, faces direct automation risk. The official downgrade of 2028 forecasts—to $50.5 billion（約8.1兆円） revenue and 2.14 million jobs—is a public acknowledgment that AI is not simply a growth opportunity for the country but a competitive threat to its primary export sector. By contrast, countries with deeper manufacturing integration (Singapore, Malaysia, Vietnam) are expected to capture incremental gains in electronics demand, though these will likely remain incremental rather than transformative.

Lee's warning that "R&D is costly" and his refusal to offer a "silver bullet solution" suggests that catching up is not simply a matter of policy willingness. Building indigenous AI research capacity and semiconductor design expertise requires sustained investment, talent, and infrastructure that take years to mature. Geopolitical fragmentation of supply chains—a theme Standard Chartered emphasized in its July 15 briefing—may create some openings for Southeast Asian diversification, but only if the region invests now in capabilities that go beyond assembly.