Gold (XAUUSD) Macro Analysis

Quarterly central bank demand has been volatile but structurally elevated throughout the observed period. From Q1 2024's 309.9t, purchases declined sharply through Q2 and Q3 2024 (211.5t and 199.5t respectively) before a decisive Q4 2024 surge to 365.1t — the strongest single quarter in the dataset. 2025 saw a sustained step-down, with Q1 through Q4 averaging 212t per quarter, reflecting a moderation from the 2024 peak cycle. Q1 2026 collapsed to just 56.5t, a 72.9% QoQ drop and the weakest quarter on record in this dataset, before Q2 2026 rebounded sharply to 288.9t — 411.3% QoQ, well above the last-4-quarter average of 195.0t. Annually, 2024's 1,092t ranks #1 in the decade, and 2025's 863t ranks #4, sitting 26% above the 5-year average of 910t — actually slightly below it but still 25% above the 10-year average of 688t. The Q1 2026 trough followed by a Q2 2026 rebound above the recent run-rate suggests the structural accumulation cycle remains intact, with the Q1 dip more likely reflecting timing and reporting lags than a genuine demand withdrawal. Central bank buying continues to provide a durable demand floor structurally above any pre-2022 baseline.

The most recent weekly flow of +22.02t (w/e 2026-08-07) marks the strongest single week in the 12-week dataset and follows +5.40t the prior week, confirming that weekly momentum is accelerating materially after a prolonged period of outflows. The 4-week net of +45.5t is sharply positive, but the 8-week net of just +8.6t reveals that this recovery is recent and built on top of heavy prior outflows — particularly the -38.22t week ending 2026-06-26, the most extreme outflow week in the series. On the monthly picture, July 2026 printed +23t versus June 2026's -74t, a significant reversal, though the YTD net of +40t, 6-month net of -80t, and 3-month net of -67t collectively tell a story of net institutional liquidation over the medium term that the most recent weeks are only beginning to reverse. The June 2026 month (-74t) and the week ending 2026-06-26 (-38.22t) stand as clear inflection-low points — ETF holders aggressively reduced exposure at that juncture, likely driven by profit-taking at or near price highs. The subsequent recovery in late July and early August, broadening geographically to include North America (+7.95t), Europe (+8.50t), and Asia (+5.66t) in the most recent week, signals that institutional re-engagement is multi-regional rather than idiosyncratic. However, the 6-month net of -80t cautions against declaring a durable trend shift — conviction is building but has not yet overcome the cumulative outflow overhang of the past three months. The flow picture is best characterised as a tentative inflection from distribution toward re-accumulation, requiring further weeks of positive momentum to confirm.

The US 10-year real yield stands at 2.39%, derived from a nominal yield of 4.63% and a 10-year breakeven of 2.24%. The mechanism behind this elevated real yield level is critical: with nominals above 4.5% and breakevens anchored near 2.2%, the real yield is high primarily because the market is pricing persistent Fed restraint rather than a collapse in inflation expectations — a distinction that matters for gold. From a macro driver perspective, nominal yields remain elevated due to Treasury supply pressure from ongoing deficit financing, a Fed holding rates restrictive longer than initially expected, and resilient labour market data reducing urgency for cuts. Breakevens near 2.24% reflect contained inflation expectations despite tariff risks, suggesting markets are not pricing a re-acceleration. Real yield is the opportunity cost of holding gold: at 2.39%, the current level is a meaningful headwind relative to the sub-1% environment that drove gold's 2020-2022 accumulation surge. However, the directional signal matters as much as the level — if nominal yields begin declining on Fed pivot expectations while breakevens hold, real yield compression would be a powerful bullish catalyst. At current levels, real yield is a net bearish input for gold on a level basis, but not prohibitively so given that structural central bank demand and de-dollarisation flows have historically overridden moderate real yield headwinds. The key variable to watch is nominal yield — any softening in US data that reprices Fed cuts would compress real yield faster than breakeven dynamics alone.

The DXY has declined 1.65 points over the most recent 30-day window, from 101.22 to 99.56 — a move that on its own appears modest but carries significant context: it represents a continuation of the broader 2026 dollar downtrend that has taken DXY from above 109 in January 2026 toward the psychologically important sub-100 threshold. The move is a clean directional trend rather than rangebound chop, consistent with sustained selling pressure rather than a one-off repricing. The primary drivers are threefold: growing fiscal credibility concerns around US deficit trajectories and foreign appetite for Treasuries, Fed policy that markets increasingly anticipate will turn more accommodative relative to a still-hawkish ECB, and episodic risk-off flows that have historically leaked out of dollar assets when US growth concerns dominate tariff-related safe-haven demand. The DXY-gold inverse relationship is structural — a weaker dollar mechanically raises gold's price in non-USD terms, broadening the global buyer base. At 99.56, the DXY is approaching sub-99 territory that is historically supportive for gold; the 103–105 range that would meaningfully pressure gold is now well above current levels. This is cyclical dollar weakness amplified by structural reserve diversification — both are bullish for gold. A sustained break below 98 on renewed Fed dovishness or US fiscal deterioration would be a powerful gold tailwind; a reversal toward 103+ on strong US data or geopolitical safe-haven dollar demand is the primary headwind. Near-term sensitivity is highest to rate differential repricing rather than structural reserve flows.