Gold is not returning to the monetary system in the form its advocates often imagine. It is unlikely to replace fiat currency, restore Bretton Woods or become the daily medium of exchange for modern economies. But it is returning in a more practical and consequential role: as reserve collateral.
Since the US severed the dollar’s convertibility to gold in 1971, fiat currencies backed by sovereign credit and confidence have dominated. That confidence is now being tested. US federal debt has crossed $39tn. Every major credit rating agency has downgraded the US from triple-A. The dollar’s share of allocated global reserves fell to 56.4% at the end of 2025 before rebounding modestly to 57.1% in early 2026. Even the world’s benchmark risk-free asset no longer feels entirely free of risk.
Central banks responded by purchasing more than 1,000 tonnes of gold annually in 2022, 2023 and 2024 – nearly triple the 2010–21 average. In 2025 that pace moderated to 863t as prices surged to successive records. But the retreat is relative, not directional: 95% of central banks surveyed by the World Gold Council still expect global official-sector gold reserves to grow over the next year, the highest reading in the survey’s eight-year history. On current trends, official world gold holdings are approaching their 1965 peak of 38,300t – a milestone that would mark gold’s return to a central, if different, role in the world monetary system half a century after it was officially retired.
The scale of that shift showed up in an unexpected place last year. According to the International Monetary Fund’s Currency Composition of Official Foreign Exchange Reserves data, gold surpassed US Treasuries as a share of official reserves in 2025 – though that development was driven almost entirely by gold’s price appreciation rather than fresh buying, and it is not reflected in COFER’s dollar-share figures, which track currencies rather than gold.
This a valuation effect, not a wave of new accumulation. But a valuation effect of that size is itself revealing. It means central banks now hold enough gold, at current prices, for its aggregate value to eclipse their Treasury holdings – a threshold that would have seemed remote a decade ago, and one that no central bank needed to vote on or announce.
I have spent much of the past year thinking about why this is happening now, and about what stands between gold’s growing monetary relevance and its actual usability. The answer to the first question is well understood: debt, deficits and the weaponisation of the dollar reserve system, most visibly through the freezing of Russian central bank assets in 2022, have made gold’s oldest virtue – that it is no one’s liability – newly valuable.
The second question is less discussed, and it is the one I want to focus on here. Gold is being rebuilt into the monetary system through balance sheets, collateral practices and market infrastructure. Gold’s problem is infrastructure, not demand.
A market voting with its actions
Central bankers rarely say the quiet part out loud, but 2024 gave us a memorable exception. Announcing a €21.6bn annual loss driven by unfavourable interest rate exposure on the Bundesbank’s government bond holdings, President Joachim Nagel was asked whether the bank might sell part of its 3,350t of gold to shore up its balance sheet. He would not consider it, he said, ‘for a nanosecond’.
The remark was more than colour: German gold, revalued at year-end prices, generated a revaluation account surplus of €193bn – nine times the surplus recorded when the euro was introduced in 1999, and more than enough to offset the bank’s quantitative-easing losses. Poland’s central bank has been even more direct about intent, lifting its gold target from 20% to 30% of reserves. Hungary, Czechia and a widening circle of central banks from the Middle East to Europe have followed similar logic, often citing the same underlying rationale: balance-sheet protection in a world where government bonds no longer behave the way reserve managers were trained to expect them to.
None of this required a change in law or a formal reset of the monetary order. It happened the way most consequential monetary shifts actually happen: quietly, asset by asset, inside institutions that have every incentive not to announce what they are doing until the pattern is undeniable.
The infrastructure problem hiding behind the demand story
This remonetisation is occurring despite an infrastructure that has barely changed since the mid-20th century. Gold custody remains fragmented across vault networks that cannot see or reconcile with one another in real time. Pricing is negotiated inside closed dealer relationships rather than discovered transparently. Settlement of institutional-size positions remains substantially manual. And the market’s own plumbing contains a structural feature that gets too little attention.
According to the London Bullion Market Association, more than 90% of interbank, wholesale and over-the-counter precious-metals trading clears through unallocated ‘Loco London’ accounts, where the holder generally has a contractual claim against the clearer rather than title to specific bars. That is not a footnote. It is a counterparty-risk layer sitting underneath an asset that institutions increasingly value because they believe it can reduce counterparty risk.
Put simply: the world is re-adopting gold as a monetary anchor faster than the market infrastructure is modernising to support that role. The practical question is how much further and faster this remonetisation could travel if the infrastructure caught up.
Redeemability, quietly reappearing
The classical gold standard rested on one mechanical promise: paper could be exchanged for gold on demand. That promise was what made the system credible. When US President Richard Nixon closed the gold window in 1971, he didn’t just change an exchange rate – he removed the redemption mechanism that had made the entire arrangement trustworthy.
Something like that mechanism is now reappearing, not through treaty but through market structure. Stablecoins were built to solve a narrower problem: giving crypto markets a stable unit of account. But the largest issuers are beginning to diversify their own reserve composition beyond Treasuries. Tether has said it aims to hold 10% to15% of its investment portfolio in gold and reportedly holds roughly 130t of bullion. That gold sits alongside equity stakes Tether has taken in gold royalty and streaming companies, a supply-side hedge that would have been unthinkable for a crypto-native firm a few years ago.
None of this makes stablecoins a gold standard. But it points towards one plausible path to a market-driven version of redeemability: if a dollar-pegged token is itself partly collateralised by gold, and the infrastructure exists to make conversion between the two seamless, auditable and instant, then holders gain something the pure fiat system stopped offering in 1971 – an exit, at their discretion, into an asset no government can print or unilaterally freeze. That is a meaningfully different proposition from simply owning gold, or simply owning a stablecoin. It is optionality, restored voluntarily by market participants rather than mandated by treaty, which is precisely the kind of gold standard that might actually survive contact with modern politics.
Separately, gold-backed tokens that attempt to circulate as everyday transactional money – PAXG and Tether’s own XAUT among them – have not achieved meaningful payment volume. The reasons are intuitive: issuers bear real storage costs, and gold’s stability – the thing that makes it valuable as collateral – makes it a poor vehicle for the kind of price movement that drives trading activity. I do not think that is where gold’s digital future lies. Gold’s evolving role looks much more like superior collateral than like a circulating medium of exchange – which is precisely why the infrastructure argument matters more than the payments argument.
What this means
To be clear, this article is not a case for a return to a rigid, treaty-based gold standard. The mechanics that made the classical system workable – a world with far less monetary demand chasing a fixed and much smaller pool of official gold – do not map cleanly onto a $39tn-debt economy with global capital mobility. Nor is this a prediction that gold will displace the dollar as the world’s primary reserve and transaction currency; the dollar’s institutional depth, liquidity and network effects remain formidable, whatever the erosion at the margin.
What I am arguing is narrower and more durable: gold is being remonetised as collateral, not currency, and that remonetisation is happening through market infrastructure rather than political decree. Central banks are demonstrating this with their balance sheets. Stablecoin issuers are beginning to demonstrate it with their reserve composition. What is missing is the connective tissue – custody, ledger, liquidity and redemption infrastructure – that would let this remonetisation proceed at the scale and speed that demand warrants.
Gold finished 2024 near record highs, reached new highs in early 2026 and has since given back a meaningful portion of that gain as markets price in higher rates and a stronger dollar. That volatility is itself instructive. A genuinely remonetised collateral asset should be expected to behave like one – priced by markets, not managed by decree, moving with the same macro forces that move every other reserve asset. The test of gold’s monetary rehabilitation was never going to be a one-way chart. It is whether, through the inevitable drawdowns, the institutions that have spent the past three years adding gold to their balance sheets keep doing so. So far, they have.
Half a century after the gold window closed, the metal is finding its way back into the architecture of financial trust – not through the front door of a formal standard, but through the side door of reserve diversification, collateral practice and, increasingly, redeemability. The task now is not to recreate a nostalgic monetary order, but to build the custody, liquidity, ledger and redemption mechanisms that a modern monetary collateral asset requires.
Steven Feldman is Co-founder and Chief Executive Officer of GBI, which owns the GBI Platform, the institutional infrastructure for physical precious metals that serves global wealth managers, banks and fintechs.
GBI is a member of OMFIF’s Gold and Precious Metals Working Group. The working group is meeting throughout the summer to discuss key questions relating to gold and its role in the global monetary system. These conversations will inform a report, publishing later this year.
Interested in this topic? Subscribe to OMFIF’s newsletter for more.