Billionaires Buy an Index Fund That Is Crushing AI Stocks Nvidia and Palantir in 2026

The SPDR Gold Shares ETF (GLD) has risen roughly 25% year-to-date and outperformed the S&P 500 by ~23 percentage points YTD and by ~52 points over the last six months, also topping recent returns from Palantir and Nvidia. Two prominent hedge fund managers — Israel Englander (added 104,900 GLD shares) and Ken Griffin (added 255,100 GLD shares and bought call options, making those calls his fourth-largest holding as of Sept. 2025) — repositioned into GLD in Q3, signaling risk-off positioning and demand for safe-haven exposure. The piece notes GLD tracks physical bullion (current quoted price $5,400/oz) and summarizes bank 2026 year-end gold targets ranging from $4,700 to $6,000 (median $5,400), while citing a roughly 11% decline in the U.S. Dollar Index during President Trump's second term as a potential driver for further gold demand.

Analysis

Market structure: The immediate winners are bullion ETFs (GLD), physical dealers, and gold miners; insurers of fiat (central banks) and long-duration bonds (TLT) could also benefit if real yields compress. Losers are high-beta equities (AI growth names like NVDA) and dollar-sensitive assets if the USD continues its ~11% YTD decline; hedge funds buying large GLD blocks and calls indicate flow-driven price amplification in the near term.

Risk assessment: Tail risks include a rapid risk-on reversal (stocks rally, USD strength) that could erase >15% of recent gold gains, coordinated central-bank selling of reserves, or a Fed surprise hike that lifts real yields by >50bp. Time horizons: days–weeks dominated by flow/option gamma (Citadel/Millennium buying), weeks–months by macro prints (CPI, employment) and Fed guidance, quarters–years by structural dollar/FX reserve shifts; watch CFTC net positioning and GLD options skew for hidden fragility.

Trade implications: Primary actionable route is allocation to GLD (cheap, liquid) and tactical call spreads to capture directional but capped exposure; miners (GDX) are leveraged alternatives but carry operational/asset risk. Cross-asset: a sustained gold rally implies falling real yields—tradeable as long TLT exposure and a tactical short on high-multiple cyclicals (relative value pair trades) while hedging with put protection on equities.

Contrarian angles: Consensus assumes continued Trump-driven turbulence; that’s one scenario but median bank targets cluster around current levels ($5,400) so upside may be narrower than momentum suggests. Historical parallels (2008/2020) show gold can drop with equities during flash deleveraging; excessive crowding into GLD/options risks fast mean reversion and liquidity squeezes — size positions accordingly and prefer defined-risk option structures.

Key Decisions for Investors

• Establish a 2–4% strategic long position in GLD (NYSEARCA: GLD) sized to portfolio risk; target $6,000/oz (≈+11%) within 6–12 months and place tactical stop-loss if GLD drops >10% from entry or if USD DXY rises >3% in 30 days.
• Deploy a defined-risk options trade: buy a 3–6 month GLD call spread (buy ~5% OTM, sell ~15% OTM) sized to 0.5–1% of portfolio notional to capture leveraged upside while capping premium bleed; unwind if GLD breaches upside leg or implied vol collapses >30% intraday.
• Enter a relative-value pair: long GLD (1.5% notional) vs short SPY (1.0% notional) for a 3–6 month hedge against risk-off; close the pair if SPY outperforms GLD by >5% over a rolling 30-day window or if 10y real yield falls >50bp (confirming direction).
• Buy a small satellite 1–2% position in GDX (gold miners) on a pullback >10% or conditional on gold >$5,700; hedge operational/alpha risk by selling short-dated miner call overwrites or buying miner puts if GLD drops >8%.
• Monitor weekly CFTC gold net-long changes, GLD options open interest (look for +50% week-over-week call OI moves), and upcoming CPI/FOMC dates — if CFTC net-longs rise sharply concurrent with large call-skew increases, trim GLD exposure by ~25% within 7 days to avoid flow-induced reversals.