Asia's K-shaped polarization is deepening: AI winners are crowded, while consumer and defensive sectors are neglected

Overview

The report uses the theme of "the best of times, the worst of times" to examine K-shaped income and consumption divergence in Japan, South Korea, China, and Taiwan, and extends this framework to Asian equity markets, sectors, and factors. The conclusion is that income inequality is driving the coexistence of premiumization and downtrading in consumption; at the market level, AI-related South Korea and Taiwan have significantly outperformed, while non-AI markets and defensive sectors have lagged; at the portfolio level, momentum, growth, and high-volatility exposures are excessively crowded, while low volatility, high dividend yield, consumer, and healthcare sectors are underowned.

Core views

The core views are: first, Asia's K-shaped economy is not a single pattern—China is the most extreme, Japan is characterized more by stagnation at the bottom, while South Korea and Taiwan show stronger gains among high-income groups with limited improvement for lower-income groups. Second, consumer investing cannot simply bet on premiumization; the real opportunities lie in experiences, emotional value, selective luxury, value-for-money offerings, and companies with exposure to global demand. Third, the K-shaped divergence in Asian equity markets is AI-led, and South Korea/Taiwan technology, momentum, and high-volatility growth are already near extremes in valuation and crowding. Fourth, if the earnings upgrade cycle peaks, the market may rotate from AI winners toward beaten-down consumer, healthcare, defensive, low-volatility, and high-dividend sectors, as well as undervalued subsegments within China technology.

Analysis framework

The report combines macro income and consumption segmentation data, industry wage differentials, country/region equity market performance, valuation indicators, earnings revisions, risk premia, sector crowding, factor performance, and sample stock screening to build a quantitative analytical framework that links economic divergence to portfolio divergence.

Methodology notes

• Macro and consumption segmentationK-shaped economy analysis

  Compare changes in income, consumption, consumption structure, and income share between the highest and lowest income quintiles.

  This framework is used to identify whether income inequality is driven by top-end growth, bottom-end stagnation, industry wage differentials, or urban-rural divergence, and to further map those drivers to opportunities in premiumization, downtrading, and experiential consumption.

• Market, sector, and factor quantK-shaped equity market decomposition

  Break market performance into three levels: country/region, sector, and factor.

  The report classifies AI-heavy markets, technology, materials, industrials, momentum, growth, and high volatility as the upward leg, and consumer, healthcare, defensives, low volatility, and high dividend yield as the downward leg, in order to assess crowded trades and potential rotation directions.

• Risk and rotation assessmentCombined framework of valuation, earnings revisions, and crowding

  Observe valuation percentiles, the earnings upgrade cycle, risk premia, correlations, and crowding simultaneously.

  When valuations and crowding are extreme, earnings upgrades are near highs, and correlations are rising, the report believes portfolios are more vulnerable to momentum reversals and style rotation shocks.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• South Korea/Taiwan AI-related technology and semiconductors

  Core assets of the upward leg

  Strengths

  Strong earnings revisions, clear theme, and support from global AI demand, with significantly leading performance year-to-date.

  Weaknesses

  Valuation, earnings expectations, concentration, and crowding are all near extremes, and market correlations are rising.

  Comparison

  They have significantly outperformed India, Indonesia, and parts of the China market, and are more crowded than defensive sectors such as consumer and healthcare.

  Risks

  If earnings upgrades peak or AI trading cools, momentum reversals and drawdowns in concentrated holdings may be triggered.

• Consumer and healthcare sectors

  Potential rotation direction that is underowned

  Strengths

  Crowding is at roughly a 20-year low, and valuation and expectation pressures have already been reflected to a considerable extent.

  Weaknesses

  There is a lack of AI-theme catalysts in the short term, and income divergence puts pressure on mass consumption.

  Comparison

  They are less popular than technology, financials, industrials, and materials, but also have greater flexibility in a style rotation.

  Risks

  If purchasing power among low-income groups remains weak, the consumer recovery could stay uneven and slow.

• Non-AI internet, entertainment, and interactive media within China technology

  Candidates for internal rotation within China tech

  Strengths

  Valuations and earnings sentiment are at relatively low levels and may benefit from rebalancing from AI hardware toward undervalued digital economy assets.

  Weaknesses

  There is still uncertainty around fundamental recovery, regulation, and consumer demand.

  Comparison

  They are cheaper and less crowded than AI hardware areas such as communications equipment, electronic equipment, and semiconductors.

  Risks

  If the market continues to reward only AI hardware, these undervalued subsegments may continue to be ignored.

• Low-volatility, high-dividend, and defensive factors

  Downward-leg assets and risk-hedging direction

  Strengths

  Current momentum is weak, and valuation and crowding pressures are low, allowing them to provide defense when risk appetite declines.

  Weaknesses

  They have clearly underperformed in a market led by AI and high-volatility growth.

  Comparison

  They are less popular than high-volatility, growth, and momentum factors, and may become the receiving end of rotation after the earnings upgrade cycle peaks.

  Risks

  If risk appetite continues to rise, defensive factors may continue to lag.

• Experiences, travel, emotional consumption, and selective luxury

  Beneficiary direction of consumption premiumization amid income divergence

  Strengths

  Consumption resilience is stronger among high-income groups, and travel, hotels and restaurants, entertainment, IP, and selective luxury goods have shown relative strength.

  Weaknesses

  Overall Chinese consumers are still affected by property and income pressures, and luxury consumption is becoming more selective.

  Comparison

  Compared with traditional mass consumption and mid-market consumption, these areas are better positioned to benefit from high-income groups and the recovery in experiential consumption.

  Risks

  If confidence among high-income consumers weakens or the travel recovery slows, sector elasticity may decline.

• Asian export-oriented and globalized companies

  Structural winning strategy

  Strengths

  They can bypass shrinking domestic consumption pools and benefit from global demand; evidence supports this in the South Korea and Taiwan AI supply chains as well as among globalized Chinese and Japanese companies.

  Weaknesses

  They are sensitive to external demand, exchange rates, trade policy, and the global technology cycle.

  Comparison

  Compared with companies relying only on domestic demand, they are better able to hedge against domestic K-shaped consumption divergence.

  Risks

  A slowdown in global demand, trade frictions, or cooling AI capital expenditure would weaken the alpha source.

Key data

• Japan income and consumption divergenceThe top 20% contribute 42.2% of income and 31.9% of spending, while the bottom 20% contribute only 6.1% of income and 10.9% of spending.Japan's K-shaped structure is mainly driven by stagnation among low-income groups; in 2025, necessities account for about 51% of spending for low-income groups versus 37% for high-income groups.
• South Korea income gapIn 2025, the highest income quintile earns about KRW11.9mn per month, versus about KRW1.3mn for the lowest quintile.Income growth at the top is faster in South Korea, while growth at the bottom is weaker; in 2025, low-income groups spend about 51% on necessities, while high-income groups spend about 40% on discretionary consumption.
• Taiwan income and industry wage differentialsIn 2024, the highest income quintile earns about NTD2.3mn annually, about 6 times the NTD384k earned by the lowest quintile.Taiwan's Gini coefficient is about 0.35, and technology manufacturing wages are about TWD95k/month, above the overall and services-sector average of about TWD63k/month.
• China income divergenceIn 2025, the top 20% have annual disposable income of about CNY103k, about 10 times that of the bottom 20%.The rural income gap is about 10x and the urban gap about 6x; differences in urban-rural conditions, access to capital, and structural opportunities intensify the K-shaped pattern.
• China urban-rural consumptionIn 2025, average urban consumption spending is about CNY36k, versus about CNY20k in rural areas.From 2019 to 2025, rural consumption spending grew at a CAGR of about 8%, versus about 5% in urban areas, but due to base effects, the absolute urban-rural consumption gap still widened at a CAGR of about 2%.
• Divergence in Asian market performanceYear-to-date, South Korea +112%, Taiwan +64%, Japan +15%, Indonesia -42%, India -10%.AI-related markets have significantly outperformed, while non-AI or weaker-AI-exposure markets have lagged, with valuations and risk premia diverging in parallel.
• Internal divergence in the China marketYear-to-date, MSCI China A +13.6%, MSCI China -12.2%.The report argues that China also has an internal K-shaped structure, with AI-related areas such as hardware and semiconductors outperforming internet, entertainment, and interactive media.
• South Korea/Taiwan technology concentrationTechnology accounts for about 85% of Taiwan's market cap and 70% of South Korea's; the top five stocks account for about 66% and 71%, respectively.Concentration, correlation, and AI theme exposure are all elevated, increasing the risk of drawdowns from a single theme.
• Factor performanceYear-to-date, Taiwan's momentum portfolio is up about +86% and South Korea's about +58%.Momentum, growth, and high-volatility factors have become the upward leg, while low volatility and high dividend yield have lagged significantly; the gap between high-volatility and low-volatility momentum within the market has reached about a 25-year high.
• Returns of export-oriented companiesFrom 2019 to 2023, Chinese companies with high overseas sales exposure and those with high overseas sales growth generated annualized alpha of 9.5% and 8.3%, respectively.The report believes that Asian companies going global and gaining exposure to global demand are a long-term winning path for coping with domestic consumption divergence.
• Japan globalization portfolioSince 2018, the globalization Japan portfolio has delivered about 5% CAGR, versus about 4% CAGR for domestic Japan; year-to-date, it is about 15% versus 4%.The report views the globalization of Japanese companies as one sustainable source of alpha.
• Travel and experiential consumptionSince 2025, South Korea entertainment and travel & leisure are up about 33%, while cosmetics are up about 27%; since 2022, Japan hotels and restaurants are at about 3% CAGR, while the consumer sector is at about -1% CAGR.Experiences, travel, emotional consumption, and selective premium consumption are relatively strong areas within K-shaped consumption.

Impact & implications

The investment implications can be divided into long-term and short-term layers. In the long term, income divergence will simultaneously support premiumization, experience/emotional consumption, selective luxury, value-for-money consumption, and Asian export-oriented companies. In the short term, AI-related winners are already highly crowded; valuations and earnings expectations for South Korea/Taiwan technology, momentum, and high-volatility growth are at extremes. If earnings upgrades slow, portfolios should consider trimming some winner exposure and increasing allocations to neglected consumer, healthcare, defensive, low-volatility, and high-dividend areas, as well as subsegments within China technology where valuations and earnings sentiment are lower.

Risks

• Valuations, crowding, and earnings expectations in AI-related South Korea/Taiwan markets are too high; if the earnings upgrade cycle peaks, a sharp rotation may occur.
• Further divergence between high-income and low-income consumption may keep mid-market mass consumption under pressure.
• Pressure from China's property market and consumer confidence may limit the breadth of recovery in premiumization and experiential consumption.
• Excessive market concentration increases portfolio dependence on a handful of technology leaders and the AI theme.
• Low-volatility, high-dividend, and defensive assets may continue to underperform if they lack catalysts while risk appetite remains strong.
• Returns of export-oriented companies are highly sensitive to global demand, exchange rates, trade policy, and the technology cycle.

What to watch

• Whether earnings revisions in South Korea and Taiwan are peaking, especially for AI hardware and semiconductor-related companies.
• Whether South Korea/Taiwan technology valuations, dividend yields, risk premia, correlations, and crowding remain at extremes.
• Whether the momentum factor spreads from high-volatility growth to low-volatility, high-dividend, and defensive factors.
• Whether internal rotation within China technology shifts from hardware and semiconductors toward undervalued areas such as internet, entertainment, and interactive media.
• Whether income and consumption structures in China, Japan, South Korea, and Taiwan continue to show the coexistence of premiumization and downtrading.
• Whether the relative returns of travel, hotels and restaurants, entertainment, IP consumption, cosmetics, and selective luxury goods continue.
• Whether overseas sales exposure and overseas sales growth of Asian export-oriented companies continue to contribute alpha.