Morgan Stanley: Asian growth retains buffers amid three pressures from AI investment, high interest rates and oil prices

Why might a slowdown in AI investment not interrupt Asia’s capital expenditure cycle?

Morgan Stanley estimates that direct AI investment will account for 7% of the increase in Asian capital expenditure over 2025–2030, and investment by semiconductor manufacturers for 5%. Even including energy and related industrial investment to power AI, the AI-related portion accounts for about 20%.
Energy, defense and industrial investment is also driven by factors such as economic security and geopolitics, and is therefore less dependent on the AI investment cycle. South Korea, Taiwan, Japan, China and Malaysia, however, are relatively more exposed to changes in AI capital expenditure.
Rising financing costs have raised concerns about returns on AI investment, with lower-rated borrowers facing greater pressure. The bank’s baseline scenario still forecasts that US AI capital expenditure in 2027 will exceed its 2026 level, even if financing, power and regulatory constraints cause delays.

So far this year, as of the report, Asian export growth has averaged 25%, above the 0.5% recorded in 2022–2023. As of the second quarter of 2026, the current account surplus had risen to 1.6 trillion dollars, or 4.2% of GDP. Exports and investment together support growth and have also improved Asia’s ability to withstand high interest rates.
Unlike the period before the Asian financial crisis in the 90s, the region has not seen a similar buildup of leverage. Corporate debt as a share of GDP in Asia excluding China remains below its pre-pandemic level.
The buffers are uneven. Morgan Stanley sees the Philippines, Indonesia and India as relatively more vulnerable to high interest rates because of their current account deficits: the deficits in Indonesia and India are both below 2% of GDP, while the Philippines’ deficit is 4.5%.

Higher export prices, especially for technology products, have improved the terms of trade and partly offset the fiscal pressure from fuel subsidies through corporate profits and tax revenues. China, India, Japan, South Korea and Thailand have sufficient refining capacity; together, the five account for 83% of regional GDP in Asia, reducing the need to import expensive refined fuels.
Government intervention has also limited increases in retail prices. Since the end of February 2026, international oil prices in local-currency terms have risen by about 60%, while domestic gasoline and diesel prices in Asia, weighted by nominal GDP in US dollars, have risen by about 19% and 21%, respectively.
The buffers come at a cost to governments or refiners. Japan, South Korea and Taiwan support refiners and wholesalers through compensation or capital injections, while Indonesia and Malaysia have increased fuel subsidies. The Philippines, by contrast, passes changes in international oil prices fully through to domestic fuel prices.

Morgan Stanley’s risk scenario combines three factors: oil prices remaining above 120 dollars per barrel, inflation forcing central banks to raise interest rates more than the bank expects, and constraints such as financing or regulation causing AI capital expenditure in 2027 to fall below its 2026 level. The scale of the shock depends on the speed, timing and duration of these factors.
Persistently high oil prices would increase the burden of subsidies, forcing more costs to be passed on to consumers, thereby raising inflation and possibly prompting central banks to raise rates while growth is under pressure. Airfare inflation in Asia is already close to its 2022 peak, and other transport prices could also rise as domestic fuel prices are raised.
The bank’s scenario analysis suggests that if oil prices reach 120 dollars per barrel or more and stay there, retail fuel prices in India could absorb half of the international increase, up from roughly one-tenth currently. Japan is also discussing raising the gasoline price ceiling month by month starting in December.