Central banks bought more gold in the second quarter of 2026 than in any quarter on record, and they did it while the price fell 16%. Both halves of that sentence matter, and the second half is the one that explains what official sector demand actually is.
This article is for informational and educational purposes only. It is not financial advice, investment advice, or a recommendation to buy, sell, or hold any security, cryptocurrency, or financial product. Always verify data with official sources before making financial decisions.
The Q2 2026 number in context
Short answer: Central banks purchased 288.9 tonnes of gold in the second quarter of 2026 — a record for gross buying, up 62% from 177.9 tonnes in the same quarter of 2025. Poland and China led. But net official sector demand, which subtracts sales, fell to 345 tonnes for the period, the weakest since 2022, because Turkey, Russia and Azerbaijan were selling heavily. Gross buying set a record while net accumulation weakened.
| Buyer | Q2 2026 purchase | Resulting holdings |
|---|---|---|
| National Bank of Poland | 51 tonnes | 632 tonnes |
| People’s Bank of China | 33 tonnes | 2,346 tonnes |
| Uzbekistan | 16 tonnes | — |
| Kazakhstan | 15 tonnes | — |
| Jordan | 6 tonnes | — |
| Czech Republic | 6 tonnes | — |
China’s 33 tonnes was its largest quarterly addition since the fourth quarter of 2023. Poland’s 51 tonnes continued a multi-year programme that has taken its reserves past 600 tonnes.
The gross-versus-net distinction
Almost all coverage of this data reports the record and omits the offset, which produces a materially misleading picture.
Gross purchases measure buying activity by accumulating central banks. Net demand subtracts sales by others. A record gross figure alongside the weakest net figure since 2022 describes a divergence within the official sector: a group of banks accumulating aggressively while another group liquidates, frequently for balance-of-payments or currency-defence reasons rather than any view on gold.
Which number is the right one depends on the question. For price impact, net is what matters — it is the change in official sector holdings that the market must absorb. For understanding reserve policy trends, gross buying by accumulators is more informative.
Buying into a falling price
The quarter’s most instructive feature is that record purchases occurred while gold fell roughly 16%.
This is not anomalous once you understand what reserve managers are doing. Central bank gold buying is a strategic allocation decision executed over years, driven by reserve composition targets rather than by price forecasts. A bank moving from 5% to 10% gold reserves will keep buying through drawdowns, and a lower price simply means the target allocation costs less to reach.
The practical implication is that official sector demand is close to price-insensitive, which makes it a stabilising source of demand and a poor signal about price direction. Central banks are not expressing a view that gold will rise; they are executing a diversification mandate.
Reserve diversification as a structural decision
The motivations reserve managers cite are consistent across surveys: gold carries no counterparty or credit risk, cannot be frozen by another jurisdiction, and has low correlation with the other assets in a reserve portfolio.
The middle point has carried more weight since 2022, when the freezing of a major central bank’s foreign reserves demonstrated that sovereign holdings of another country’s debt are contingent on that country’s cooperation. Gold held domestically is not. This is the substance behind the de-dollarisation framing, and it is narrower than that term suggests — it is a risk management response to a specific demonstrated risk, not a wholesale rejection of the dollar system.
Reading the World Gold Council release
The quarterly Gold Demand Trends report is the standard source and repays careful reading. Total demand is split across jewellery, technology, investment and official sector, and the official sector line is where central bank activity sits.
Two cautions. Reported figures rely on what central banks disclose to the IMF, and disclosure practices vary — some report with substantial delay, and analysts have long suspected undisclosed accumulation by certain buyers. And “central bank demand” is not always a central bank: sovereign wealth funds and state entities are treated differently across sources.
Risks, uncertainty, and limits
Reported purchase figures are revised as additional disclosures arrive, and quarterly numbers frequently change in later publications. The gap between reported and estimated buying is real and unquantifiable.
Nothing here forecasts the gold price. Record official sector buying coincided with a 16% quarterly decline, which is the clearest possible demonstration that this data is not a price signal.