We're booking some big profits in one stock and buying more shares in another

Selling 20 shares of Goldman Sachs at about $866 to reduce the Trust's GS weight to 4.65% from 5.11% (realizing ~55% gain on Dec 2024 purchases) and moving the rating to a hold-equivalent. Using proceeds to buy 50 shares of Alphabet at roughly $298, lifting GOOGL holdings to 350 shares and weight to 3% from 2.57% to lower average cost; rationale cites Alphabet's AI leadership and analyst note estimating a $2.5B/$7.5B incremental Google Cloud revenue opportunity in 2026/2027. Trades are portfolio rebalancing actions rather than changes in fundamental thesis; individual-stock impact is modest.

Analysis

Rebalancing away from idiosyncratic financial exposure into AI/compute leaders materially shifts portfolio risk from near-term event volatility to multi-year feature adoption and infrastructure monetization. That tilt increases sensitivity to capital-intensive, lumpy revenue streams (compute licensing, cloud infrastructure) which typically manifest in GAAP flows 12–24 months after contract announcements, so realized upside will lag sentiment moves.

The Broadcom-to-TPU dynamic is a structural accelerant for an outsourced compute stack: it props up third-party silicon vendors and server OEMs while compressing the pure-play GPU vendor narrative if specialized ASICs gain share. This causes a subtle winners/losers bifurcation — platform owners with proprietary data and end-to-end stacks see widening moats, whereas advertising-dependent social platforms face outsized downside from litigation and attention-risk shocks.

Primary tail risks are clear — a banking shock or macro liquidity squeeze could reprice cyclicals and trigger correlation breakdowns within weeks, while tech-specific risks (litigation, antitrust, failed licensing rollouts) can reverse momentum over 3–18 months. Monitor three catalysts as timing signals: material licensing revenue disclosures (6–18 months), next major earnings season for banks (near-term volatility), and regulatory/legal verdicts that alter ad or compute economics.

Given the cautious backdrop, trading should be conviction-weighted and hedged: favor asymmetric option structures to capture multi-year upside without blowing up on near-term drawdowns. Maintain strict stop rules on event-driven exposures and tilt size toward names that monetize compute with recurring, high-margin contracts rather than advertising-facing growth that is litigation-sensitive.

Key Decisions for Investors

• Initiate a core long GOOGL position (1–2% NAV) with a 12–24 month horizon; complement with a Jan-2027 call (buy) to create convex upside. Target objective: 30–50% upside if compute licensing/AI monetization accelerates; hard stop: reduce exposure by 50% on a 15% drawdown from entry.
• Establish a pair: long GOOGL / short META at a 1:0.5 notional ratio for 3–12 months to exploit differential regulatory and litigation sensitivity. Expect asymmetric payoff if compute monetization narrative outperforms ad-engagement recovery; cap downside by sizing short to 30–50% of long notional.
• Buy 3–9 month protective puts on a small financials sleeve (GS or a regional bank ETF) instead of outright selling equities to hedge event risk around bank volatility; aim to limit drawdown from sudden credit/liquidity repricing while keeping optionality for recovery. Cost should be <1% NAV for a tactical hedge.
• Small tactical long in AVGO via a 12-month call spread to play third-party silicon and licensing optionality (buy lower strike, sell higher strike). This captures upside if licensing scales while capping premium outlay and limiting single-stock vega risk.