Central banks around the world are quietly stockpiling four times more gold than usual and are moving it out of the U.S.

Goldman Sachs estimates central banks bought 79 tonnes of gold in August, more than four times the pre-2022 average of 17 tonnes; seasonally adjusted purchases were running at 85 tonnes per month over three months. China accounted for an estimated 25 tonnes, while central banks are also shifting storage locations: the Dutch central bank moved 86 tonnes from New York and Ottawa to London, and Banque de France sold 129 tonnes in New York and bought gold in Europe. Goldman reported a 40-tonne decline in New York Fed holdings and a 46-tonne rise at the Bank of England in August, alongside growing reluctance to disclose storage plans.

Analysis

The investable signal is a more persistent, less price-sensitive source of gold demand—not evidence that the dollar’s reserve role is about to break. If reserve managers keep diversifying, that can support gold through risk-off episodes and reduce the metal’s sensitivity to any single geopolitical headline. But custody transfers change where bullion is available, not the global quantity of gold: any London-versus-New York dislocation should be treated as a potential liquidity or basis issue, not an automatic shortage thesis.

The key contrarian point is measurement risk. Purchases are partly inferred, undisclosed buyers may be material, and one month’s estimate should not be annualized. The market may be underpricing the durability of official demand, while overreading it as a near-term catalyst; real yields, the dollar and investor flows can still dominate short-term pricing. For the next 1–3 months, watch central-bank disclosures, gold ETF flows, the dollar and real yields. Over 6–18 months, continued opaque diversification would strengthen the structural floor, but a reversal in reserve behavior or a sustained rise in real yields would weaken it.

Key Decisions for Investors

• Consider a measured long in gold through GLD or IAU on pullbacks rather than chasing geopolitical spikes. The thesis is persistent official-sector demand; the principal near-term falsifiers are a sustained rise in real yields, a stronger dollar alongside weakening gold ETF flows, or evidence that estimated central-bank buying is overstated.
• Treat the custody shift as a watch item, not a standalone trade. Monitor London–New York spreads and lease rates for persistent dislocations; without confirmation in those market indicators, relocation alone does not establish a tradable scarcity premium.
• Do not extrapolate the monthly purchase estimate into a fixed demand run-rate. Reassess after subsequent central-bank data and disclosures; if buying estimates roll over while real yields rise, reduce gold exposure rather than assuming reserve diversification will support price regardless of macro conditions.