Central Bank Gold Reserves 2026: The New Global Accumulation Cycle

For the third consecutive year, central banks collectively purchased more than 1,000 tonnes of gold in 2025 — a pace of accumulation unseen since the collapse of Bretton Woods. The People's Bank of China alone added an estimated 280 tonnes, while the National Bank of Poland, the Reserve Bank of India, and the Monetary Authority of Singapore each bought at levels that would have been considered extraordinary a decade ago. This is not a cyclical blip. It represents a structural shift in how sovereign reserve managers think about gold — and it has profound implications for the gold market through the end of this decade.

The Scale of the Buying: By the Numbers

The World Gold Council's central bank survey data, corroborated by IMF International Financial Statistics and individual central bank disclosures, paints a clear picture. Central bank net purchases totalled 1,037 tonnes in 2024 and an estimated 1,089 tonnes in 2025, following 1,082 tonnes in 2023. Prior to 2022, the post-2010 average was roughly 470 tonnes per year. The current cycle is running at more than double the long-run trend.

The composition of buying has shifted meaningfully. During the 2010–2020 cycle, emerging-market central banks — China, Russia, Kazakhstan, Turkey — accounted for roughly 80% of net purchases. The current cycle is broader. In 2025, Poland, India, Singapore, and the Czech Republic joined the traditional EM buyers, while even developed-market central banks with historically static gold holdings — including the Banque de France and the Deutsche Bundesbank — signalled in public statements that they had no intention of reducing gold allocations and viewed the metal as a strategic reserve asset.

What's Driving the Accumulation: Five Structural Factors

The narrative explanation — "geopolitical uncertainty" — is directionally correct but analytically insufficient. Central bank reserve managers operate within formal frameworks with explicit objectives: capital preservation, liquidity, and (for some) return optimisation. The shift toward gold reflects concrete changes in how these frameworks assess the risk profile of traditional reserve assets.

1. Sanctions Risk and Reserve Freezing

The freezing of approximately $300 billion in Russian central bank reserves held in G7 jurisdictions in 2022 was a watershed moment for reserve managers globally. Whatever one's view of the merits, the operational precedent was unambiguous: sovereign reserves held in foreign government securities and central bank deposit accounts are not immune to political risk. This was previously understood in theory but never demonstrated at scale against a G20 economy.

The implication for reserve composition is straightforward. Gold held in a central bank's own vault — or in allocated storage at the Bank of England, the Federal Reserve Bank of New York, or the Banque de France — is not subject to counterparty freeze risk in the way that electronic ledger entries representing Treasury or Bund holdings are. For reserve managers at the People's Bank of China, the Reserve Bank of India, and central banks across the Global South, this distinction has moved from an academic consideration to a core portfolio-construction parameter.

2. The Decline of the Dollar Share

The dollar's share of global allocated reserves has declined from roughly 71% in 2001 to approximately 57% in Q4 2025, according to the IMF's Currency Composition of Official Foreign Exchange Reserves (COFER) data. This is not a dramatic collapse — the dollar remains the dominant reserve currency by an enormous margin — but it is a persistent, multi-decade trend that reserve managers are responding to.

The composition of the shift is instructive. The declining dollar share has not been replaced by a rising euro, yen, or renminbi share in proportionate amounts. Instead, a growing portion of reserves is flowing into "other" — a category that includes gold, but also smaller reserve currencies (Australian dollar, Canadian dollar, Korean won, Swedish krona) and non-traditional reserve assets. Gold is the largest and most liquid component of this "other" category, and it benefits disproportionately from the diversification flow.

3. Sovereign Debt Sustainability Concerns

The fiscal trajectory of the United States — with federal debt-to-GDP exceeding 120% and annual deficits running above 6% of GDP — has introduced a new variable into reserve manager calculations. Central banks hold US Treasuries not because they are bullish on American fiscal policy, but because the Treasury market is the only market deep enough to absorb sovereign reserve flows at scale. But as the stock of outstanding debt grows faster than GDP, the long-run credit risk, however remote, becomes a factor that cannot be entirely ignored.

Gold, by contrast, has no issuer and no credit risk. It is nobody's liability. In a world where the supply of "risk-free" sovereign debt is growing faster than the economies that back it, an asset with zero credit risk and a fixed (or slowly growing) above-ground stock has a compelling structural bid from investors whose primary mandate is capital preservation.

4. Portfolio Mathematics: Gold's Role in Reserve Optimisation

Beyond the geopolitical narrative, there is a cold mathematical case for higher gold allocations in sovereign reserve portfolios. A 2025 working paper by economists at the Bank for International Settlements modelled optimal reserve composition under a range of scenarios and found that gold allocations of 8–15% of total reserves were optimal for most central banks under current conditions — well above the global average of roughly 5% (and below 3% for many Asian central banks excluding China).

The drivers are: (a) gold's near-zero correlation with both equity and fixed-income markets over multi-year horizons, (b) its strong negative correlation with the trade-weighted dollar, which benefits non-US central banks whose liabilities are in local currency, and (c) its convex payoff profile during tail-risk events — gold tends to rally precisely when other reserve assets are under stress. Mean-variance optimisation, Black-Litterman models, and CVaR frameworks all point in the same direction: current gold allocations are sub-optimal for most central banks.

5. The BRICS Payment System and Gold-Backed Settlement

While the BRICS grouping's ambitions for a common currency have receded from the headlines, the quieter work of building alternative payment infrastructure continues. The BRICS Bridge platform — a messaging and settlement system developed as an alternative to SWIFT — entered pilot phase in late 2025 with participation from China, Russia, India, Brazil, and South Africa. While not a "gold-backed currency" in any meaningful sense, the platform incorporates gold as a settlement asset for netting imbalances, giving participating central banks an operational reason to hold physical gold beyond the standard reserve-diversification logic.

This development should not be overstated — SWIFT remains dominant, and BRICS Bridge volumes are negligible in global terms. But for the central banks involved, the existence of a gold-settlement mechanism creates a marginal incentive to accumulate that did not previously exist, and it provides a policy rationale for purchases that might otherwise be questioned by domestic constituencies.

The Supply-Demand Arithmetic

The demand side of the gold market is well understood. What is less frequently analysed is the interaction between central bank demand and the physical supply constraints of the gold mining industry.

Global mine production has been essentially flat since 2018 at approximately 3,600–3,650 tonnes per year. New discoveries are declining in both frequency and grade. The average grade of gold deposits discovered in the past decade is roughly half that of deposits discovered in the 1990s. Meanwhile, all-in sustaining costs (AISC) have risen from approximately $900/oz in 2016 to roughly $1,400/oz in 2025, driven by higher energy costs, labour inflation, and the increasing proportion of production from lower-grade, more remote deposits.

The implication: annual mine supply of roughly 3,650 tonnes must satisfy total demand of approximately 4,700 tonnes (jewellery ~2,000t, central banks ~1,090t, bars and coins ~1,100t, industrial/technology ~320t, ETFs and similar ~190t). The shortfall — roughly 1,050 tonnes per year — is met by recycling (scrap gold), which is itself price-elastic and sensitive to economic conditions in major scrap-supplying countries like India and China.

If central bank demand persists at or above 1,000 tonnes per year — and the structural drivers outlined above suggest it will — the physical market will remain in a structural deficit that must be cleared by higher prices to either reduce price-elastic jewellery and bar demand or increase scrap supply. Neither mechanism works on short timescales; both imply a higher equilibrium gold price over the medium term.

China: The Strategic Imperative

The People's Bank of China deserves separate treatment, both because of the scale of its buying and because its motivations are distinct from those of other large purchasers. China's reported gold reserves stood at approximately 2,400 tonnes at end-2025, up from roughly 1,950 tonnes at end-2022. This represents an increase of approximately 23% in three years.

Several factors specific to China are at work:

• Reserve composition imbalance: China holds approximately $3.2 trillion in foreign exchange reserves, of which gold constitutes roughly 6% by value — well below the 70%+ for the United States and 65%+ for the Eurozone. If China were to target even a 15% gold allocation, it would need to acquire an additional 5,000+ tonnes at current prices — more than an entire year's global mine production.
• De-dollarisation strategy: The PBOC's gold accumulation is part of a broader effort to reduce exposure to dollar-denominated assets, which includes diversifying into non-dollar reserve currencies, promoting renminbi-denominated trade settlement, and developing the Shanghai Gold Exchange as an alternative price-discovery venue to the LBMA/COMEX complex.
• Domestic gold market development: China is both the world's largest gold producer (approximately 370 tonnes/year) and its largest consumer. The PBOC's purchases support the development of Shanghai as a gold trading hub and reinforce the renminbi's credibility as a gold-backed-adjacent currency in international trade.

It is noteworthy that the PBOC's reported purchases — which are disclosed to the IMF with a lag and are widely suspected to understate actual buying — resumed in November 2022 after a reporting hiatus of more than two years. The timing, coinciding with the post-invasion sanctions regime and the freezing of Russian reserves, was unlikely to be coincidental.

Poland and India: The New Heavyweights

While China's buying commands the headlines, the purchases by the National Bank of Poland (NBP) and the Reserve Bank of India (RBI) are in some respects more significant, because they reflect decisions by democratic governments with transparent policy frameworks — not opaque strategic imperatives.

Poland's NBP has been one of the most aggressive buyers globally, with Governor Adam Glapinski publicly stating a target of 20% of reserves in gold. At end-2025, Poland held approximately 450 tonnes, up from 230 tonnes at end-2022. The NBP's rationale is explicitly about tail-risk insurance: Poland's geographic position on NATO's eastern flank, combined with its history, makes reserve security a matter of genuine strategic concern rather than a theoretical portfolio exercise. The NBP has also repatriated a significant portion of its gold from the Bank of England to its own vaults in Warsaw — a decision that reflects the same sanctions-risk calculus discussed above.

India's RBI added approximately 95 tonnes in 2025, bringing its total to roughly 900 tonnes. India's gold accumulation is driven by a different logic: the desire to diversify a reserve portfolio that remains heavily concentrated in dollar-denominated assets, combined with a cultural and institutional comfort with gold that most other central banks lack. The RBI also repatriated over 100 tonnes of gold from the Bank of England in 2024 — the largest single repatriation by any central bank in recent memory — citing "operational flexibility" and "diversification of storage locations."

Implications for the Gold Market Through 2030

If the structural drivers identified above persist — and we see no reason they will not — central bank demand is likely to remain in the 900–1,200 tonne per year range through the end of the decade. Combined with flat mine supply and relatively inelastic jewellery and bar demand, this implies a market that clears at progressively higher prices.

Several specific implications follow:

• A higher floor: Central bank buying establishes a price-insensitive bid that did not exist at this scale in previous cycles. During the 2013–2015 gold bear market, central banks were buying roughly 400–600 tonnes per year and the price fell from $1,900 to $1,050. With central bank demand now running at roughly double those levels, the price floor is structurally higher — our trading desk estimates a floor of approximately $2,200–$2,400/oz under current conditions, versus roughly $1,200–$1,400 in the previous cycle.
• Reduced volatility: Central banks are long-term, strategic buyers. They do not trade in and out based on technical signals or short-term rate expectations. Their presence as a consistent source of demand absorbs selling pressure during corrections, dampening downside volatility. The 30-day realised volatility of gold has trended lower over the past three years even as the price has risen — an unusual combination that reflects this structural bid.
• Decoupling from real yields: Historically, gold's price exhibited a strong inverse relationship with US real yields (TIPS). That relationship has weakened materially since 2022 as central bank buying — which is not sensitive to the opportunity cost of holding gold — has become the dominant marginal demand driver. Real yields rose sharply in 2023–2025; gold rose alongside them. This decoupling has confounded macro funds positioned for a gold selloff on higher real rates and suggests that traditional valuation frameworks need updating.
• Mining equity re-rating potential: If the gold price floor is structurally higher, gold mining equities — which have traded at a persistent discount to the metal itself since the 2013 bear market — could experience a sustained re-rating. Producers with long-life, low-cost assets in stable jurisdictions are the primary beneficiaries. The GDX and GDXJ indices have begun to reflect this dynamic but remain below their 2011 inflation-adjusted highs.

Risks to the Thesis

No investment thesis is complete without a clear-eyed assessment of what could go wrong. The central bank accumulation narrative faces several risks:

• Geopolitical de-escalation: A meaningful reduction in geopolitical tensions — particularly a resolution to the Russia-Ukraine conflict and a stabilisation of US-China relations — would weaken the sanctions-risk and reserve-diversification drivers, potentially reducing central bank buying toward pre-2022 levels.
• Fiscal consolidation in the US: A credible medium-term fiscal consolidation plan that stabilises or reduces the US debt-to-GDP ratio would strengthen the dollar and reduce the reserve-diversification impetus. This seems unlikely in the current political environment, but it cannot be ruled out over a five-year horizon.
• Central bank selling: While buying attracts attention, the risk of coordinated or large-scale central bank selling — as occurred under the Central Bank Gold Agreement (CBGA) framework between 1999 and 2019 — cannot be entirely dismissed. The CBGA signatories have not sold meaningful volumes since 2020, but the framework technically expired rather than being permanently abolished.
• Technological substitution: The emergence of central bank digital currencies (CBDCs) and blockchain-based settlement systems could, over the very long term, reduce the need for physical gold as a settlement asset by providing alternative mechanisms for cross-border payment finality without counterparty risk. This is a 2030s risk, not a 2026 risk, but it warrants monitoring.

What This Means for Institutional Investors

Central bank buying is not directly investable — you cannot replicate the PBOC's position or ride the RBI's coattails. But the second-order effects are significant for institutional portfolios:

• Higher structural allocation to gold: If central banks — the most conservative reserve managers in the world — are increasing gold allocations, institutional investors with similar mandates (endowments, sovereign wealth funds, family offices) should at minimum re-examine their own gold allocations. The old 2–5% allocation rule of thumb may be obsolete in a world where the marginal buyer is adding hundreds of tonnes per year.
• Physical over synthetic: The sanctions-risk driver that motivates central banks to favour physical gold over paper claims applies with equal force to institutional investors. Allocated physical gold in a non-bank vault in a jurisdiction with strong property rights provides insurance against financial system disruption that a GLD ETF share or a COMEX futures contract cannot replicate. The premium for physical delivery over paper settlement has widened in recent years and may continue to do so.
• Producer equities as a leveraged play: For investors comfortable with equity risk, gold mining equities offer leveraged exposure to a structurally higher gold price. The operational leverage inherent in mining — where a $100/oz increase in the gold price flows almost entirely to free cash flow once AISC is covered — means that producers' earnings and share prices can appreciate by multiples of the underlying gold price move.

The Bottom Line

Central bank gold demand at or above 1,000 tonnes per year is not a temporary phenomenon driven by a single geopolitical event. It reflects a structural reassessment of reserve-asset risk that has been underway since 2022 and shows no sign of reversing. The sanctions precedent, the declining dollar share of reserves, sovereign debt sustainability concerns, and the mathematical case for higher gold allocations in optimised portfolios all point in the same direction: central bank buying is likely to remain elevated through the end of the decade.

For the gold market, this means a structurally higher floor, reduced volatility, and a sustained supply-demand deficit that will clear through higher prices. For institutional investors, it means gold deserves a larger allocation than most portfolios currently give it — and that physical gold, held outside the banking system, offers insurance characteristics that paper substitutes cannot match.

The central banks have spoken, not through speeches, but through their balance sheets. The question for institutional investors is whether to listen.