Institutional Investors Are Selling One of Wall Street's Premier Trillion-Dollar Club Members (No, Not Nvidia!)

Institutional holdings in Taiwan Semiconductor Manufacturing fell 2.8% in the December quarter to ~789.6 million shares, while institutional ownership rose for other trillion-dollar names including Nvidia, Broadcom, Berkshire Hathaway, Eli Lilly and Walmart. The TSMC selling appears driven largely by profit-taking after a jump from ~$220–240 to $290–310, with passive/index funds doing much of the rebalancing and hedge funds only marginally reducing exposure. Trade uncertainty tied to President Trump’s tariff/trade policy is cited as an additional possible catalyst for institutional exits.

Analysis

Index-driven and flow volatility in large-cap semiconductors creates transient dislocations that can misprice structural secular demand for AI compute. Foundries sit at the intersection of a capital-intensive supply response and concentrated customer power; a modest slowdown in incremental capex (6–12 months) would amplify lead-time shortages 12–24 months out, benefitting suppliers with secured capacity and hurting those running spot-dependent procurement models.

Geopolitical and trade policy uncertainty acts as a multiplier on the capex cycle: even a 1–2 quarter delay in capacity expansion raises effective utilization by mid-teens percentage points, which pushes ASPs higher for constrained nodes. Near-term catalysts that could reverse sentiment include large order confirmations from hyperscalers, new onshore fab announcements with firm financing, or a visible inventory draw at OEMs — each has a distinct timing window (days–weeks for order prints, months for capex announcements, quarters for inventory signals).

Competitive second-order winners are equipment vendors and specialty logic/analog players with multi-sourced supply (e.g., equipment vendors capturing outsized aftermarket revenue when utilization spikes). Losers are companies overexposed to single-point foundry risk or to passive flow volatility; their multiples can compress faster than fundamentals move, creating tactical shorts or pair opportunities.

The consensus underweights the probability that supply-side underinvestment translates into pricing power for leading-node suppliers over 12–24 months. That creates a two-phase trade: exploit short-term flow-driven weakness with hedged exposure to the structural bull case, rather than outright directional bets unprotected against geopolitical tail risk.

Key Decisions for Investors

• Long NVDA via a 3–6 month call spread (buy 25% OTM, sell 60% OTM) sized as 2–4% of portfolio — target asymmetric 2.5x payoff if AI order cadence accelerates; stop-loss if NVDA falls 12% intraday from entry due to flow unwind.
• Pair trade: Long AVGO (6–12 month horizon) and short TSM (6–12 month horizon) size 1:1 notional — AVGO is less concentrated geopolitically and captures system-level content gains; expect 20–35% upside on AVGO vs 15–25% downside protection on the short via buying a 9–12 month 15% OTM put on TSM to cap tail risk.
• Contrarian accumulation: Dollar-cost average TSM on 10–20% drawdowns with a 12–24 month hold, hedged by buying 9–12 month puts at ~20% OTM to protect versus geopolitical shocks; risk/reward aims for ~25–40% IRR if secular capacity tightness materializes.
• Volatility play: Buy 60–90 day straddles/strangles on major policy or earnings windows for TSM to capture event-driven IV spikes — allocate small (0.5–1% portfolio) due to theta decay, take profits on IV expansion or close before last-week decay.