J.P. Morgan: Taiwan is the Core of AI Profits, Recommends TWSE Call Option Spreads

Overview

This report by J.P. Morgan's Asia Pacific derivatives strategy team argues that Taiwan equities (TWSE) have become the most concentrated and direct beneficiary of the current AI profit cycle. The report notes that as the 4-5 earnings season confirms AI trends remain hardware-driven, Taiwan's absolute advantage in advanced manufacturing and packaging bottlenecks translates to superior earnings revision breadth versus global peers. Combined with foreign capital inflows and upcoming MSCI passive allocations, the report recommends leveraging TWSE's relatively low implied volatility window to capture AI hardware upside through call option spreads or cross-market Worst-of Calls. The report also details Gamma position changes and fund flow divergences in NKY, KOSPI2, and HSCEI.

Core views

The AI profit cycle remains hardware-centric, with Taiwan occupying the 'bottleneck' niche. A dual-axis exposure framework reveals TWSE's AI hardware weight at 72%, significantly higher than KOSPI2's 63% and NDX. Crucially, this exposure is concentrated in supply-constrained segments like advanced wafer fabrication and packaging, rather than generic hardware participation. This structural advantage directly translates to earnings quality: TWSE's overall earnings revision breadth is 53%, with AI hardware constituents at 77%, far ahead of NDX (62%) and KOSPI2 (36%). In contrast, Hong Kong tech indices are more exposed to AI software/applications, exhibiting weaker elasticity in the current hardware-driven cycle. Fund flows and technicals converge, with MSCI rebalancing as a near-term catalyst. Despite TWSE's 40.5% YTD rally, flows show a positive reversal—foreign investors net bought $7.8 billion in April after ~$30 billion outflows in March. More critically, EM ex-China fund inflows are accelerating, with Taiwan representing ~1/3 of the index. MSCI's free-float methodology change effective post-May 29, 2026, is expected to bring passive inflows equivalent to ~103% of TSMC's ADTV, adding technical support to fundamental strength. Derivatives pricing mismatch presents entry opportunities. Despite supportive fundamentals and flows, TWSE's 2-month ATM implied volatility lags KOSPI2 and TSMC single-stock levels. The report proposes specific trades: a TWSE Jun-2026 105%-115% call spread (priced at 2.19%, Delta 26%) balances cost control with upside capture. For broader regional AI hardware exposure at lower premium, a TWSE/KOSPI2/NKY Worst-of Call (Jul-2026 105% strike) costs just 2.1%, ~50% of vanilla call pricing. Divergent derivatives microstructure across Japan/Korea/HK. NKY's break above 60,000 deepened dealer short gamma to ~$1.7 billion (7% of futures turnover), concentrated in puts as investors hedge profits—likely easing post-May 8 expiry. KOSPI2's record highs see dealer short gamma contained at ~$191 million, but record leveraged ETF AUM poses risks—a 10% move could trigger ~$5 billion in dealer hedging flows (15% of 3M futures ADV). Korean single-stock structured products shifted focus from overseas tech to local Samsung/SK Hynix. HSCEI oscillates in 8600-8900 range with dealer positions flipping to long (~$415 million), reflecting weak call demand and defensive sentiment.

Analysis framework

The report employs a four-dimensional framework: 'fundamental positioning + earnings verification + flow confirmation + derivatives pricing.' First, it avoids simplistic sector classifications, instead constructing a 'AI hardware vs. AI applications' dual-axis model to pinpoint Taiwan's unique bottleneck positioning. Second, high-frequency 'earnings revision breadth' metrics validate earnings season signals, distinguishing hardware/software performance delivery. Third, macro EM ex-China flows are combined with micro MSCI rule changes to quantify passive flow impact. Finally, implied volatility cross-comparisons and Gamma positioning monitor trading mismatches. This closed-loop analysis from industrial logic to execution exemplifies derivatives strategy research.

Methodology notes

• Industry/Industrial Analysis Framework

  AI Supply Chain Dual-Axis Exposure Framework (Hardware vs. Applications)

  Instead of traditional GICS classifications, the report categorizes index constituents by their actual AI value chain roles—'AI hardware' (e.g., chip manufacturing, infrastructure) and 'AI applications' (e.g., internet platforms, software services). This better identifies beneficiaries of current capex cycles, avoiding misclassification of concept stocks without tangible earnings.

• Company Fundamentals & Financial FrameworkEarnings Quality Analysis

  Earnings Revision Breadth

  This metric measures the proportion of companies with upward earnings revisions over the past month, reflecting the breadth and persistence of improvements more effectively than average EPS revisions. The report uses it to demonstrate the universality of Taiwan AI hardware earnings upgrades, enhancing confidence in index-level rallies.

• Quantitative/Factor/Portfolio TheoryBeta/alpha analysis

  Gamma Positioning & Market Volatility Feedback Mechanism

  Gamma measures dealers' hedging sensitivity to market moves. When dealers are 'short gamma,' market declines force additional futures selling, exacerbating downturns—and vice versa. By monitoring Gamma net positions, the report assesses whether markets are in 'pro-cyclical' unstable states or 'mean-reverting' stable states to evaluate technical risks.

• Event Gaming & Behavioral FinanceEvent-driven analysis

  Passive Flows Triggered by Index Rule Changes

  Benchmark index rule adjustments (e.g., MSCI free-float methodology changes) compel passive funds to rebalance. These flows are mechanically deterministic, creating concentrated buying/selling pressure. The report treats them as non-fundamental technical catalysts for timing or yield enhancement.

Asset mapping & comparison

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

• TSMC (2330 TT)

  Core AI hardware bottleneck beneficiary, direct MSCI rebalancing winner

  Strengths

  Monopoly in advanced nodes, leading earnings revision breadth, deterministic passive inflows

  Comparison

  Versus Samsung/SK Hynix, TSMC holds higher barriers and stronger earnings visibility in AI compute manufacturing

  Risks

  Geopolitical risks, customer concentration

• Samsung Electronics / SK Hynix

  Korean AI hardware representatives, focus of local structured product shifts

  Strengths

  HBM leadership, domestic structured product demand support

  Weaknesses

  AI hardware weight (63%) trails Taiwan, earnings revision breadth (36%) lags significantly

  Comparison

  As KOSPI2's AI hardware anchors, their earnings elasticity and certainty trail TSMC

  Risks

  Gamma risks from oversized leveraged ETFs, retail leverage

• TWSE Index

  Optimal index-level proxy for AI hardware profit cycle

  Strengths

  72% AI hardware weight, 77% revision breadth, foreign inflows + MSCI catalysts, low implied vol

  Weaknesses

  40.5% YTD rally elevates valuations

  Comparison

  Versus NDX, KOSPI2, NKY, TWSE leads in AI hardware purity and earnings delivery

  Risks

  Foreign flow reversals, MSCI rebalancing priced in early

Key data

• TWSE YTD Gain40.5%As of report date, validates prior bullish TSMC earnings view
• TWSE AI Hardware Weight72%Significantly higher than KOSPI2 (63%) and NDX, highest among major Asian indices
• TWSE AI Hardware Earnings Revision Breadth77%Past month data, far exceeds NDX (62%) and KOSPI2 (36%)
• April Foreign Net Buying in Taiwan Stocks$7.8 billionReverses ~$30 billion March outflows
• MSCI Impact Estimate for TSMC~103% ADTVPassive inflows from post-May 29 free-float adjustments
• TWSE 105%-115% Call Spread Pricing2.19%Jun-2026 expiry, IV 31.7v/30.5v, Delta 26%
• NKY Dealer Short Gamma Estimate~$1.7 billion~7% of futures turnover, concentrated in puts
• KOSPI2 Leveraged ETF Potential Hedging Flows~$5 billionTriggered by 10% volatility move, 15% of 3M futures ADV
• Korea April Structured Product Issuance$1.7 billionMonthly YTD high, single-stock focus shifted to Samsung/SK Hynix

Impact & implications

The report argues that Asia Pacific AI trends have entered a 'survival of the fittest' phase, where only hardware firms at supply bottlenecks sustain earnings upgrades. Taiwan equities, given their unique industrial structure, offer the most efficient exposure. For derivatives investors, TWSE's low volatility provides a rare 'cheap calls' window—both single-name spreads and cross-market combos offer better cost efficiency than outright equity or vanilla calls. Meanwhile, risks loom from Korean leveraged ETFs and NKY short gamma amplifying non-fundamental volatility. Hong Kong remains peripheral in the current AI narrative due to its lack of hardware focus and defensive skew.

Risks

• Record Korean leveraged ETF AUM could trigger massive dealer hedging flows during volatility spikes, exacerbating overshoots
• NKY's concentrated short gamma in near-month puts may amplify downside moves during earnings season or BOJ policy shifts
• MSCI free-float adjustment passive inflows may be front-run, diluting actual impact
• Retail margin loans and deposits rebounding to highs elevate Korean market leverage risks
• Periodic geopolitical tensions may suppress HK and Asia tech risk appetite

What to watch

• Post-May 29 MSCI free-float adjustment actual passive inflows into TSMC
• AUM growth and Gamma impact of newly listed Korean single-stock leveraged ETFs (Samsung/SK Hynix)
• Whether NKY's post-May 8 expiry alleviates dealer short gamma as expected
• Whether TWSE implied volatility rises rapidly during rallies, eroding option strategy cost advantages
• Diverging foreign flow rhythms across Taiwan/Korea/Japan cash and derivatives markets