Hedge funds’ selling of emerging Asian market stocks last week was most in a year, says Goldman

Global hedge funds executed the heaviest selling of emerging Asia stocks since April 2025, with the selloff concentrated in Taiwan, Korea and India and driven mainly by short-selling. Taiwan's benchmark plunged ~5% and Korean stocks fell ~3% in early Monday trade after escalating US-Iran threats; all major regions were net sold, led by North America and emerging Asia. Despite the outflows, hedge funds' exposure to emerging Asia remains near record highs, and Korea/Taiwan have been top YTD performers on semiconductor and AI-led demand (Samsung, SK Hynix, TSMC).

Analysis

Concentrated positioning in semiconductor-related names has become a market-structure vulnerability: when risk-off shocks hit, liquidity evaporates fastest in the same handful of high-beta tech stocks, amplifying drawdowns and skewing realized volatility 2-3x the index for short bursts. The characteristic is asymmetric—downside cascades quickly on forced deleveraging, while upside requires a conviction-driven, lower-liquidity re-entry that can produce sharp short-covering snap-backs.

From a competitive angle, companies with US-listed, vertically integrated exposure to AI hardware demand (high-margin appliance and systems vendors) will re-rate differently from foundry and wafer-makers domiciled in geopolitically sensitive jurisdictions. That divergence is not just headline risk: supply-chain friction raises effective capex lead times and insurance/shipping premiums, widening gross margins for vendors that can source domestically or pass through costs.

Time horizons matter. In the next 3-10 trading days, monitor positioning-driven flows and FX moves—these are the likeliest triggers for another 5-12% swing in individual EM/tech names. Over 3-12 months, fundamentals (inventory digestion, AI-driven revenue upgrades, and any diplomatic de-escalation) will determine whether current repricings are a buying window or a regime shift to structurally higher risk premia.

The consensus trade (de-risking EM semis) understates convexity: if volatility mean-reverts, a narrow basket of high-operating-leverage AI plays can outperform materially as shorts scramble to cover. Execution should therefore combine defined-risk bullish exposure to US-listed AI hardware beneficiaries with selective, time-boxed protection on Taiwan/Korea-domiciled equities to capture that asymmetric payoff.

Key Decisions for Investors

• Buy SMCI 3-6 month call spread (buy ~25-delta, sell ~10-delta) to capture continued AI-driven demand while capping premium; target 40-60% return on premium, max loss = premium paid. Entry: on 5-10% intraday weakness; stop: widen >20% realized vol spike.
• Pair trade — long SMCI / short TSM (size to be dollar neutral) over 3 months to express US-listed AI hardware outperformance vs Taiwan-foundry geopolitical discount; target 20-30% relative outperformance, stop if pair moves against by 15% (reassess on de-escalation news).
• Buy TSM 2-4 month put spread (buy 25-delta put, sell 10-delta put) to hedge tail geopolitical risk while limiting cost; target: protection for a 10-20% downside move, cost ≈ small fraction of notional, max loss = premium paid.
• Use APP as a liquidity-friendly thematic long: buy APP 6-12 month LEAP calls (30-40% notional of a concentrated AI hardware position) as a lower-volatility growth exposure to offset higher-beta system names; position size should be scaled to limit portfolio delta to neutral.
• Monitor GS volatility and flow signals intraday for tactical entry/exit (use GS-provided flow analytics if available); if hedge-fund gross exposure shows further unwind in a single week, trim directional long exposure by 25% and redeploy into the pair/defined-risk structures above within 24-72 hours.