AI Boom Drives Record-High Current Account Surpluses in South Korea and Taiwan

Overview

This report provides an in-depth analysis of the profound impact of the Artificial Intelligence (AI) boom on the macroeconomics of South Korea and Taiwan. It states that the current AI surge represents the strongest technology cycle in the history of both nations, driving a spike in AI-related exports and pushing current account surpluses to record highs. Despite pressures from dependency on Middle Eastern energy imports and weakness in non-tech sectors, the scale of tech exports is sufficient to offset these negative factors. The report focuses on how surpluses bypass domestic economies for recycling abroad and the resulting monetary policy and FX implications, forecasting that the KRW will perform stronger relative to the TWD, and that the Bank of Korea will commence a rate hiking cycle in the second half of 2026.

Core views

AI-driven Export Boom: Sales of AI-related technologies (such as chips, servers, and High Bandwidth Memory HBM) are experiencing unprecedented growth. In South Korea, AI-related memory sales are expected to triple in 2026, reaching nearly 30% of GDP (less than 10% over the past five years); in Taiwan, AI-related tech sales are projected to exceed 30% of GDP. This growth is primarily driven by the shift from training workloads to inference and agentic AI workloads, increasing demand for high-end memory. Current Account Super Surpluses: A surge in tech exports will drive current account surpluses to new highs. In 2026, South Korea's current account surplus is expected to far exceed 10% of GDP, and Taiwan's to surpass 20%. Through a "Chip-Energy Balance" analysis, the report indicates that even under an adverse scenario of severe Middle Eastern energy price shocks, the surplus would remain intact because the scale and growth of chip exports completely dominate the energy price path. It estimates that the 2026 Chip-Energy trade surplus could reach 25% of GDP for South Korea and approximately 28% for Taiwan. Differentiated Paths for Surplus Recycling: The large current account surpluses are currently mostly bypassing domestic economies, failing to significantly boost broad money (M2 growth remains around 5%) or immediately lead to sharp currency appreciation. There is a divergence in how surpluses are handled in both regions: South Korea's surpluses flow primarily through households and pension funds into overseas equity investments, whereas Taiwan's surpluses are converted mainly into overseas deposits (accumulating nearly 8% of GDP in a single year in 2025). This difference means South Korea faces greater volatility in financial flows, while Taiwan accumulates a massive offshore fund pool. Divergence in Macro Growth and Inflation: Both economies are expected to grow strongly in 2026, with South Korea's GDP growth rebounding from 1.0% in 2025 to 2.5%, and Taiwan accelerating from 8.7% to nearly 10%. The tech sector is the main driver, contributing most of the real GDP growth. However, continued weakness in non-tech sectors (accounting for about 90% of South Korea's GDP and 75% of Taiwan's) limits labor market pressure and overall domestic demand. Regarding inflation, South Korea faces higher upside risks, mainly stemming from wage negotiations in the tech sector (potentially increasing wage compensation share of GDP from 0.5% to 1.75%) and pressure from rising housing prices; Taiwan's tech sector wages remain relatively stable, and the central bank's macro-prudential tools (such as loan-to-value ratio restrictions) effectively curb overheating in the housing market. Policy and FX Outlook: Given the "K-shaped" cycle characteristics (strong tech, weak non-tech), the report advocates for targeted and prudent fiscal policy, saving a portion of excess tech income to mitigate pro-cyclicality. On monetary policy, it expects the Central Bank of Taiwan to raise rates by 12.5bp in Q2 and Q4 (cumulative 25bp); for South Korea, the view is adjusted from "hold" to two hikes totaling 50bp in the second half of the year, with a terminal rate of 3.0%. On FX, both face appreciation pressures, but the KRW may outperform the TWD due to a larger improvement in external balance, a more flexible exchange rate regime, regulatory changes aimed at improving net financial flows, and greater room for policy rate hikes.

Analysis framework

The report adopts a combined approach of macro supply-demand frameworks and industry chain transmission analysis. First, by decomposing the AI technology cycle (from training to inference/agentic AI), it quantifies the contribution of exports to GDP through demand pull in specific semiconductor segments like HBM. Second, it introduces a specific analytical indicator, "Chip-Energy Balance," to assess the resilience of tech export surpluses under volatile energy import costs, thereby isolating the interference of energy prices from traditional current account analysis. Furthermore, the report contrasts the surplus recycling mechanisms in the financial accounts of both countries (equity investment vs. overseas deposits), and combines differences in exchange rate regimes (free floating vs. managed float) and monetary policy frameworks (inflation targeting vs. money growth targeting) to explain why Taiwan's financial markets react more moderately to positive growth surprises. This progressive analysis, moving from real trade to financial flows and then to policy responses, clearly delineates the transmission path of AI surpluses to macro variables.

Methodology notes

• Industry/Industrial Analysis FrameworkSupply-demand framework

  AI Technology Iteration Drives Structural Changes in Semiconductor Demand

  The report infers a structural increase in demand for specific hardware like High Bandwidth Memory (HBM) by analyzing the shift in AI workloads from "training" to "inference" and "agentic AI," representing a typical logic of deriving upstream supply needs from downstream application scenarios.

• Macroeconomic framework

  Chip-Energy Balance Analysis

  The report constructs a specific macro indicator—the ratio of net semiconductor exports minus energy import costs to GDP—to filter out the interference of energy price fluctuations and more purely measure the real contribution capability of tech exports to the balance of payments.

• Macroeconomic framework

  Surplus Recycling Mechanism Analysis

  By analyzing how current account surpluses are transformed into financial account outflows (such as purchasing overseas stocks or accumulating overseas deposits), the report evaluates their actual impact on domestic money supply (M2) and exchange rates, revealing transmission breakpoints between real trade surpluses and asset allocation.

Key data

• Share of 2026 South Korea AI-related Exports in GDPNearly 30%Significant increase from less than 10% over the past decade
• Share of 2026 Taiwan AI-related Tech Sales in GDPOver 30%Further increase following strong growth in 2025
• 2026 Forecast for South Korea Current Account SurplusExceeds 10% of GDPRecord high
• 2026 Forecast for Taiwan Current Account SurplusExceeds 20% of GDPRecord high
• 2026 Forecast for South Korea GDP Growth2.5%Significant rebound from 1.0% in 2025
• 2026 Forecast for Taiwan GDP GrowthNearly 10%Acceleration from 8.7% in 2025
• 2026 Interest Rate Hike Expectation for Bank of KoreaTwo 25bp hikesReport view shifts from hold to hike, terminal rate 3.0%
• 2026 Interest Rate Hike Expectation for Central Bank of TaiwanCumulative 25bp12.5bp in Q2 and Q4 respectively

Impact & implications

The report posits that AI-driven super surpluses will reshape the macro landscape for both South Korea and Taiwan. For investors, this implies that strong performance in the tech sector will continue to support economic growth in both countries, but also brings increased risk of growth volatility (especially in Taiwan). On the policy front, central banks will not monetize external surpluses but will instead address potential inflation and exchange rate pressures through rate hikes. In the FX market, the KRW may show stronger appreciation potential than the TWD due to more thorough improvements in fundamentals and greater room for policy adjustments. Additionally, the link between South Korea's tech cycle and the real estate market could bring additional inflation and financial stability risks, warranting close monitoring of wage negotiation outcomes.

Risks

• High volatility in the global tech cycle could lead to unexpectedly large swings in growth
• If wage negotiation results in South Korea's tech sector exceed expectations significantly, it could drive up inflation and housing prices
• Extreme unfavorable movements in Middle Eastern energy prices, while having limited impact, still constitute a headwind
• Taiwan's enormous scale of overseas deposits without repatriation measures could affect the efficiency of exchange rate management

What to watch

• Wage negotiation outcomes for South Korea's two major memory manufacturers
• Effectiveness of Taiwan's central bank macro-prudential policies (such as loan-to-value ratio restrictions)
• Changes in South Korea's regulatory policies regarding overseas equity investments and their impact on capital outflows
• Sustainability of demand for high-end memory driven by AI inference and agentic AI workloads