Quick Take
- Central banks bought 321 tons of gold in Q3, exceeding jewelry consumption.
- About 63% of estimated official buying wasn't publicly reported.
- Gold and Bitcoin both fell in October, despite their appeal as scarce assets.
Central banks bought more gold in the third quarter of 2026 than consumers worldwide bought for jewelry.
According to the World Gold Council's latest demand report, central banks and other official institutions acquired an estimated net 321 metric tons, while jewelry consumption came to 313 tons.
In the same quarter a year earlier, jewelry buyers had consumed about 150 tons more than central banks acquired.
By September, the institutions buying gold for their reserves had overtaken the people buying it to wear, although roughly 63% of estimated official demand couldn't be traced through publicly reported country-level purchases.
The price barely kept pace with Bitcoin, which gained 42.7% during the quarter versus gold's 3.84%. Both attracted money through exchange-traded funds, so this wasn't a straightforward case of investors ditching one for the other. They bought both and got different results.
However, Bitcoin's lead came crashing down in October: gold fell 6.3%, while Bitcoin lost 7.3%. Governments kept accumulating bullion, and investors kept believing in scarce assets, but neither belief did much for that month's returns.
The buyers behind those beliefs, and their reasons for holding on, explain rather more than the safe-haven label does.
The $188 billion gold rush that wasn't
Global gold demand reached $188 billion in the third quarter, up 22% from a year earlier. That sounds like an impressive shopping spree until you weigh the purchases.
Buyers took home 1% less metal, at 1,369 tons, while the average quarterly price was 23% higher. Much of the extra spending covered the price increase rather than buying more gold.
That was especially apparent in jewelry, where consumption fell 17% to 313 tons, the weakest third quarter in the council's records since 2000. Buyers in China and India cut purchases or chose lighter and lower-carat pieces, adjusting what they bought to what they could afford.
Physical investment wasn't immune either: demand for bars and coins fell 19% to 268 tons. Paying more per ounce didn't make every buyer more eager to own one.
Central banks' net purchases rose 42% to 321 tons, while gold-backed ETFs added 211 tons, reversing the previous quarter's withdrawals.
Those purchases helped offset weaker demand from jewelry buyers and physical investors, giving the market support from buyers who wanted gold as a financial holding even as others balked at the cost.
The categories don't divide neatly into institutions buying and households retreating. Individuals own gold through ETFs, and institutions can own bars and coins.
But the numbers do show that demand weakened in traditional consumer and physical-investment markets, while official reserves and listed funds absorbed more metal. Gold was becoming a harder purchase to justify at the counter and an easier allocation to defend in a portfolio.
Of the estimated 321 tons in net purchases, only about 120 tons appeared in publicly reported country-level activity.
China disclosed additions of 63 tons and Poland 40 tons, making them the largest reported buyers, but the country disclosures leave approximately 201 tons of the council's estimate unaccounted for in the public figures.
Reporting delays and purchases governments don't disclose can explain activity missing from the country data, and the broader estimate may be revised. The gap alone also doesn't establish secret, coordinated purchases.
We have a much better estimate of the total than a complete account of who bought it, which limits how confidently we can interpret the buying.
Q3 2026 net official demand, using rounded inputs: 321 metric tons estimated in total, of which 120 were publicly reported. The remaining 201 are calculated by subtraction, not added to the total. Unreported activity can include delayed or undisclosed purchases; estimates may be revised.
- Publicly reported net buying120 metric tons37.4%
- Estimated unreported net buying201 metric tons62.6%
CryptoSlateCentral banks acquired an estimated 675 tons in the first nine months of 2026, exceeding their purchases over the same period of 2025.
Their appetite is large enough to offset weakness elsewhere, but incomplete disclosures make it harder to attribute that appetite to particular governments or motives. The survey evidence offers more detail on the latter, with some important limits of its own.
Quarterly totals. Jewelry measures consumption; bars and coins, ETFs, and official institutions measure net demand. Negative ETF values mean net withdrawals. Official institutions include central banks. These selected channels do not exhaust total gold demand and do not separate household from institutional investors.
CryptoSlateReserve managers are rethinking the dollar, but don't have to sell it
In the World Gold Council's 2026 central-bank survey, 74% of respondents expected the US dollar's share of global reserves to be lower in five years. That gives some basis for the argument that reserve managers want less dependence on the dollar.
The IMF's second-quarter foreign-exchange reserve data put the dollar's share at 56.7%, down from 57.18% in the preceding quarter. It remained the dominant reserve currency, with dollar holdings largely steady as holdings in other currencies expanded.
This series also excludes monetary gold, so it can't tell us how much gold replaced dollars.
Even the funding of gold purchases is less straightforward than the familiar sell-Treasuries-buy-bullion argument allows. Half the survey respondents said they'd finance new gold purchases through domestic programs using local currency, while 38% said they'd sell existing reserve assets.
Buying locally produced gold with local currency doesn't require a corresponding sale of US government debt, and the quarterly demand totals can't identify what financed every purchase.
Reserve managers can want more gold without wanting out of everything else. Bullion gives them an asset whose value doesn't depend on an issuer honoring a financial obligation, making it a way to reduce reliance on other governments' currencies and debt.
However, the bargain still includes a market price, and that price can fall enough to make the word “protection” feel rather generous over shorter periods.
Gold finished October at $3,913 an ounce, down 6.3% for the month and roughly 25% below its January peak, according to the World Gold Council's November market commentary.
Central bank demand hadn't spared holders that loss, because even a persistent buyer can't dictate the price everyone else trades at. Futures positions and the need to raise cash can overwhelm the reassurance of a government adding bullion to its reserves.
Global gold-backed ETFs attracted $2.7 billion in net inflows during October, adding 19 tons and taking holdings to a record 4,275 tons. Their combined assets under management still fell 6% to $542 billion, as the lower gold price more than offset the new money.
The funds held more metal and were worth less, an entirely possible outcome that gets lost when inflows are treated as a verdict on returns.
The broader point is consistent with a July 2026 IMF study, which found that gold's ability to hedge market risk depends on the shock and the economic conditions around it. Gold can help diversify a reserve portfolio over time without reliably sheltering it during extreme stress.
Someone reducing exposure to sovereign credit over several years and someone trying to avoid a loss this month may buy the same asset, but they're asking it to do different jobs.
Bitcoin outran gold while buyers backed both
Bitcoin gave investors a much better third quarter, gaining 42.7% to end September around $83,570, according to 21Shares' quarterly analysis.
US spot Bitcoin ETFs added about 79,800 BTC, worth about $6.2 billion, reversing roughly $4.8 billion in second-quarter withdrawals.
Their purchases were nearly twice the estimated 41,300 BTC newly mined over the period, giving the rally a measurable source of demand through funds available to professional and individual investors alike.
That comparison with new supply is striking, although newly mined coins are only one source of Bitcoin available to buy. Existing holders own a far larger stock, and some were reducing their positions while ETFs accumulated.
21Shares estimated that balances in the long-term-holder cohort fell by 248,000 BTC during the quarter. The figure isn't a direct count of exchange sales, since coins can leave the cohort through transfers or other classification effects, but it prevents ETF buying from being mistaken for a market in which everyone was adding to their holdings.
Global gold-backed funds attracted a record $31 billion in third-quarter inflows and added 211 tons of bullion, even as the metal's price gain fell well short of Bitcoin's.
The dollar totals cover different markets: the gold figure is global, while the Bitcoin figure covers US spot ETFs. Their respective market sizes and investor bases differ too, so treating the totals as a head-to-head contest would tell us little.
Investors added exposure to both assets, but their quarterly returns differed by nearly 39%, which offers little support for a simple account of Bitcoin rallying because money was leaving gold. It also limits what ETF data alone can explain.
Listed funds supplied demand in both markets, while trading elsewhere, including derivatives positions and borrowed money, helped determine how far each price moved.
October exposed the difference between holding an asset and hiding in it
October narrowed that enormous performance gap in an unwelcome way. Bitcoin fell 7.3% to about $77,500, while gold lost 6.3% over the same month.
Reported October 2026 returns in US dollars. Gold uses the LBMA PM benchmark through Oct. 30, the last business day; Bitcoin trades continuously. These are price returns, excluding fees and income, rather than ETF returns. Similar losses in one month do not establish a lasting correlation or a common cause.
CryptoSlateGold ETFs continued to attract net investment globally, led by Europe and Asia, while North American funds recorded modest withdrawals. On Nov. 4, Reuters reported that the largest Bitcoin-backed funds tracked by LSEG/Lipper lost $215 million in October.
While the totals don't establish that investors moved money from Bitcoin to gold, they show that net buying persisted across global bullion funds while the tracked Bitcoin funds lost money to withdrawals.
The World Gold Council pointed to reductions in speculative positions, revised expectations for Federal Reserve policy, and rotation between assets as possible contributors to gold's decline.
Bitcoin contended with weaker tech stocks, tighter credit conditions, the Binance hack, and the US midterm elections. Different sellers had different reasons to act, even though holders of both assets ended the month with losses.
Scarcity offers little defense against someone who needs cash or wants less exposure to volatile markets. Central-bank purchases can absorb gold without guaranteeing its price, just as Bitcoin's supply limit can constrain issuance without guaranteeing buyers at the price its holders would prefer.
October put both propositions on display, even while some investors continued to add bullion to their portfolios.
Governments holding Bitcoin aren't the same as central banks buying it
The sovereign comparison needs a careful approach, because a government owning Bitcoin doesn't necessarily mean it has entered the market to buy it.
The US established a Strategic Bitcoin Reserve in March 2025, formally giving Bitcoin a place among assets the federal government can retain and manage.
However, the executive order creating the reserve directed that it be funded primarily with Bitcoin already held through criminal or civil asset forfeitures.
Officials were also authorized to devise budget-neutral ways to acquire additional Bitcoin without incremental taxpayer costs. That permission and the decision to retain existing coins have policy importance, but neither establishes a recurring purchase program comparable with central banks adding gold to their reserves.
That's why putting seized Bitcoin into a reserve can happen without buying a single additional coin on the market.
The World Gold Council's quarterly figures estimate net additions to official gold holdings. The US reserve order governs, in large part, what happens to assets the government already owns.
Gold already has an established role in central bank balance sheets and reserve management. Bitcoin's newer sovereign role could evolve, but ETF access, government recognition, and official purchases each affect demand through a different route.
Investors can't assume progress on one will produce the buying associated with another, however attractive the reserve announcement sounds.
Liquidity will test what scarcity can promise
Interest-rate expectations and access to cash will continue to test both markets. Stronger US employment or inflation readings can encourage expectations of tighter Federal Reserve policy, increasing the appeal of interest-bearing assets relative to gold and potentially making financial conditions less accommodating for Bitcoin.
Weaker readings can support expectations of easier policy, although a deterioration severe enough to prompt fears of an economic slowdown can also leave investors wanting cash.
Neither the gold-demand report nor Bitcoin's ETF inflows can settle how investors will react to the next release. They can tell us more about who has been buying and how readily those buyers might leave.
Central banks can accumulate gold over several years, while ETF investors can cut exposure much faster. Consumers can buy lighter jewelry, and traders using borrowed money can face demands to reduce positions precisely when they'd prefer to wait.
That mix helps explain why persistent belief in an asset can coexist with a sharp fall in its price.
The reserve manager diversifying away from another government's obligations may have little reason to sell bullion during a correction, while an investor facing losses elsewhere may sell the same metal because it's readily tradable.
Bitcoin's fixed supply doesn't remove that need for cash, and official recognition doesn't give its holders the resources or mandate to ride out every downturn.
Central banks overtaking jewelry buyers shows how gold demand is being redistributed, and October's losses show the limits of what those purchases can promise other holders.
Someone buying for protection needs the asset's behavior to fit the risk they're trying to reduce and the period they can afford to hold it.
Gold can serve a reserve manager well over years and disappoint an investor over weeks, while Bitcoin can deliver a spectacular quarter without becoming a dependable place to wait out the next sell-off.
Mentioned in this article
Editorial credits
