Asia/EM equity market allocation and AI/industrial capex themes Report Interpretation

Overview

This Asia/EM equity-allocation report reassesses the region after a severe momentum and leverage unwind. Morgan Stanley remains constructive on AI infrastructure and the industrial super-cycle, upgrades Korea and Thailand, and uses Australia as an underweight funding source while retaining Japan and Singapore as core overweights.

Core views

Morgan Stanley argues that the recent Asia/EM correction was primarily technical rather than a break in the AI and industrial super-cycle thesis. The APxJ 12x1-month price-momentum factor fell about 30%, its third-largest drawdown since 2002, while Korea’s local momentum factor fell roughly 40%. The institution sees a better entry point into AI computing infrastructure and related industrial themes, but expects leadership to broaden in 2H26 rather than remain concentrated in the earlier winners. The Korea upgrade rests on cheaper valuation, cleaner positioning and sustained thematic earnings support. Morgan Stanley sees 36% upside to its Korean strategist’s KOSPI target of 9,000, with a near-term range of 5,500–10,500. KOSPI trades at 5.7x 12-month forward P/E; excluding Samsung Electronics and SK hynix, it trades at 10.8x, down from a 14.8x peak. Consensus forecasts MSCI Korea EPS growth of 312% in 2026 and a further 20% EPS CAGR in 2027–28. The report argues that this valuation already discounts a major reversal in memory earnings, while Korea also offers exposure to AI infrastructure, power, energy security, defence, renewables and humanoids. The report treats the leverage washout as evidence that forced selling is increasingly exhausted, though not a guarantee of an immediate bottom. Asian hedge-fund de-grossing is estimated to be roughly 75% complete; gross exposure to Korea fell to 2.7% from 3.8% and net exposure to 5.6% from 8.8%. Leveraged ETF AUM linked to Samsung Electronics and SK hynix fell about 70%, from US$40 billion in June to US$11.8 billion on July 29, while short gamma declined 67% from US$867 million to US$288 million. Korea’s Capitulation Index reached -2.53, its lowest reading since 2008 other than crisis episodes, and 30-day realized volatility reached an all-time high. Morgan Stanley notes that past sharp volatility spikes were followed by 10–30% rallies over the next 30 trading days, but says foreign inflows are needed for the next leg higher because retail margin balances, though down 14%, remain elevated. Morgan Stanley remains constructive on AI spending and demand despite debate over hyperscaler capex, utilization and Chinese open-weight models. Its base case is that computing remains constrained and that physical and political limits on US data-centre development create speed bumps rather than barriers; it cites a projected 38GW US data-centre power shortfall in 2026–28 and a potential capacity shortfall above 10% by 2028. Its US Internet team raised 2027 and 2028 hyperscaler capex estimates by 9% and 10%, respectively, to about US$1.2 trillion and US$1.4 trillion. The report estimates that US$2–5 of token cost can produce US$55 of savings in a typical enterprise use case, cites productivity gains above 10% for firms with more than 12 months of AI implementation, and expects 65% of companies to have AI/LLM models in production by end-2026, versus 22% at end-2025. It also argues that Chinese memory supply is not a near-term HBM solution because yields, technology gaps and lithography access limit effective output. For Korean chipmakers, the report highlights capital-management announcements, HBM4 pricing and the September iPhone 18 launch as catalysts. HBM4 could represent around 60% of Samsung’s HBM sales by the end of the following year; pricing above US$3/Gb would support DRAM pricing in 2026–27. Mobile demand remains 30–40% of global DRAM demand and 25–30% of NAND demand, and the report cites potential 5–10% year-on-year iPhone 18 volume upside. At the portfolio level, Morgan Stanley favors a Korean barbell led by Industrials, followed by Banks and Tech, with Communication Services, Healthcare and Staples also preferred to broaden exposure and moderate volatility. Thailand is upgraded on an improved post-election policy backdrop, stronger FDI, structural reform and inexpensive valuation. The Bank of Thailand has cut its policy rate to 1%, and Morgan Stanley expects low nominal rates alongside expansionary fiscal policy, including a deficit around 4% of GDP, to reduce real rates and support reflation. FDI is around 3.3% of GDP, close to a 10-year high, and the government’s reform agenda targets potential growth above 3% and a top-20 global competitiveness ranking by 2030. Thailand received US$881 million of foreign equity inflows in 1H26 and US$211 million in June. MSCI Thailand’s median stock trades at 14.1x forward P/E, the 10th percentile of its 10-year range, with a 4.2% dividend yield near the top of its historical range. The report also sees Thailand as an emerging data-centre and energy-security beneficiary, supported by about 10GW of spare baseload power capacity and a projected 3GW of data-centre capacity by 2035, up from a prior 1GW forecast. Australia is moved to underweight because Morgan Stanley expects tighter monetary and fiscal conditions, housing-tax reform and a softer domestic economy to limit index upside. The RBA has tightened 75bp in 2026 after 75bp of cuts in 2025, while the FY27 budget is the first contractionary fiscal package since FY18. Changes to negative gearing and capital-gains taxation alter property-investment returns; the team estimates a 15–20% housing-price fall would restore investor returns to equilibrium, though its base case is a 5–10% national adjustment. The report expects aggregate EPS growth to compress from low double digits to mid single digits. The ASX 200 trades at 17.4x forward P/E versus a 14.9x long-term average, and Morgan Stanley’s 9,250 target implies limited upside. Within Australia, it prefers Resources and capex-exposed companies to domestic cyclicals, citing projected capex growth from A$225 billion in FY25 to A$326 billion in FY30. Elsewhere, Morgan Stanley maintains core overweights on Japan and Singapore, citing reform agendas and YTD returns of 11% and 19%, respectively. It remains underweight Indonesia and the Philippines, stays equal weight across much of North Asia ex-Japan, and favors Taiwan’s technology supply chain and industrial-super-cycle beneficiaries. The report expects Taiwan and Japan to show the strongest earnings-preview breadth, sees Australia, Latin America and China/HK facing continued downgrades, and identifies India as a potential earnings-inflexion market. It also expects broader market breadth and greater diversification away from concentrated technology leadership through the rest of 2026.

Analysis framework

Morgan Stanley combines relative country allocation with earnings revisions, forward valuation, profitability, policy and sector-theme analysis. It cross-checks the fundamental case with flows, hedge-fund and long-only positioning, leverage, momentum, volatility and breadth indicators, then uses earnings-preview screens and correlation analysis to identify areas where a fundamental earnings inflection may support broader leadership.

Methodology notes

• Valuation methodsPB valuation

  Forward P/B and ROE comparison

  The report compares markets’ forward price-to-book multiples with forward return on equity and changes since 2022 to judge whether profitability improvements are reflected in valuations.

• Quantitative, Factor, and Portfolio TheoryStyle factor analysis

  Momentum, Value and Quality factor analysis

  Morgan Stanley measures the magnitude of the momentum unwind and observes that Value and Quality outperformed during Korea’s correction, using this to assess style rotation and market breadth.

• Event-Driven and Behavioral FinanceFund-Flow and Positioning Analysis

  Positioning, leverage, flow and capitulation indicators

  Hedge-fund exposure, leveraged ETF AUM, short gamma, retail margin balances and foreign flows are used to assess whether forced selling and crowding have eased.

Asset mapping & comparison

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

• Korea / KOSPI

  Upgraded market allocation supported by valuation, earnings and cleaner positioning.

  Strengths

  36% upside to the 9,000 KOSPI target; 5.7x forward P/E; exposure to AI infrastructure and industrial themes.

  Weaknesses

  High concentration in Samsung Electronics and SK hynix.

  Comparison

  Morgan Stanley prefers Korea over Australia after the correction.

  Risks

  Retail leverage remains elevated and renewed foreign inflows are needed for further upside.

• Thailand / MSCI Thailand

  Upgraded market allocation based on accommodative policy, FDI, reforms and valuation.

  Strengths

  14.1x forward P/E at the 10th percentile of 10-year history; FDI near a decade high; AI data-centre and energy-security exposure.

  Weaknesses

  High household debt remains a structural constraint.

  Comparison

  The report contrasts Thailand’s improving policy direction with Australia’s tightening backdrop.

  Risks

  Energy dependence, higher fuel prices, early policy tightening and weak reform execution.

• Australia / ASX 200

  Underweight funding source in the regional allocation.

  Strengths

  Resources and capex-exposed sectors retain relative appeal within Australia.

  Weaknesses

  Tighter policy, housing-tax changes and deteriorating domestic earnings outlook.

  Comparison

  Morgan Stanley favors Japan, Singapore, Korea and Thailand over Australia at the index level.

  Risks

  Persistent inflation and restrictive policy underpin the 6,850 bear-case target.

• Gulf Development (GULF.BK)

  Added to APxJ and GEM Focus Lists as a Thai AI-power and infrastructure beneficiary.

  Strengths

  Integrated LNG, power and data-centre platform; 1GW data-centre ambition by 2030; about 11% 2025–29E core earnings CAGR; 80% of capacity contracted under PPAs with fuel pass-through.

  Comparison

  Presented as a key Thai hyperscaler-infrastructure opportunity.

• Samsung Electronics and SK hynix

  Major Korean AI and memory exposures supporting Korea’s valuation case.

  Strengths

  HBM and memory-cycle exposure; leverage-linked positioning has materially reset.

  Weaknesses

  Their 49% KOSPI weight can overshadow broader index earnings.

  Comparison

  The report sees Chinese memory producers as more relevant to commodity DRAM/NAND than to near-term HBM supply.

  Risks

  AI capex and ROIC uncertainty, Chinese memory progress, and dependence on capital-management catalysts.

Key data

• KOSPI target upside36% to 9,000Morgan Stanley’s Korean strategist target; near-term range is 5,500–10,500.
• KOSPI valuation5.7x 12-month forward P/EEx-technology valuation is 10.8x, down from a 14.8x peak.
• MSCI Korea EPS growth312% in 2026; 20% CAGR in 2027–28Consensus forecasts cited by the report.
• Korea leveraged-ETF AUMUS$11.8bnDown about 70% from the US$40bn June peak as of July 29, 2026.
• Thailand valuation14.1x 12-month forward P/EAt the 10th percentile of the past 10 years; median dividend yield is 4.2%.
• Thailand equity inflowsUS$881mn in 1H26; US$211mn in JuneSecond-largest Asian recipient in 1H26 and the top net-inflow destination in June.
• Australia valuation17.4x 12-month forward P/EAbove the 14.9x long-term average.
• Hyperscaler capex forecast~US$1.2tr in 2027 and ~US$1.4tr in 2028Estimates were raised 9% and 10%, respectively.

Impact & implications

The report favors reallocating within Asia/EM rather than abandoning AI-related and industrial-capex themes. It sees Korea as a discounted, cleaner-positioned route to several structural themes, Thailand as a policy-and-valuation recovery opportunity, and Australia as a less attractive source of domestic cyclicals risk; it also argues for broader sector exposure as market concentration eases.

Risks

• Weak third-quarter seasonality, long-term-yield-driven valuation pressure and higher energy prices remain region-wide risks.
• US data-centre development faces political, power, environmental and community constraints.
• Korean upside depends on renewed foreign investor inflows while retail margin balances remain elevated.
• Thailand faces structural energy dependence, higher fuel-price pressure, possible premature policy tightening and uncertainty over reform execution.
• Australia faces persistent inflation, restrictive policy, housing-price weakness and earnings-delivery risk.

What to watch

• AI capex and return-on-investment evidence, including hyperscaler spending and enterprise AI adoption.
• Korean chipmaker capital-management announcements, HBM4 pricing and the September iPhone 18 launch.
• Foreign flows into Korea, retail margin balances, hedge-fund exposure and market volatility.
• Thailand’s Fast Pass and OECD-linked reform implementation, policy accommodation and FDI trends.
• Australian inflation, RBA policy, housing-market adjustment and the effect of FY27 tax reforms on banks and consumers.