Gold’s current rally is rooted in a sequential expansion of structurally distinct buyer classes, sovereign, institutional, and crypto-native, each adding demand without displacing prior layers, in contrast to prior gold cycles where price strength depended on a single dominant category of buyer.
- Central banks purchased above 1,000 tonnes annually for three consecutive years (2022–2024), establishing a sovereign demand floor that preceded the return of Western investment flows.
- ETF and private capital re-entered in 2025, adding 801 tonnes, but Western portfolios remain under-allocated at 0.17% of U.S. private financial assets versus a historical norm closer to 1–2%, leaving significant room for expansion without requiring new buyer categories.
- Crypto-native demand is forming a structurally independent third layer through tokenized gold (35–40t, $6B+), stablecoin reserves (Tether $20B), and yield-bearing structures, introducing gold as productive collateral rather than passive reserve in digital financial systems.
Common characterisations of gold as a macro hedge describe what gold does in portfolios; they do not identify who has been buying, in what size, or why that buying has persisted across three years of rising prices.
The defining feature of this gold cycle is the sequencing of buyers. Three overlapping demand phases have developed independently.
Phase 1: The Sovereign Floor (2022–2024)
Central banks established the foundation of this cycle before ETF flows or retail participation returned in any meaningful size. Annual net purchases exceeded 1,000 tonnes for three consecutive years between 2022 and 2024, the prior single-year record was around 610 tonnes in 2013. The scale and persistence of this accumulation was without modern precedent.
The structural characteristics of central-bank demand explain why this created a durable floor rather than a temporary spike. Reserve allocation is strategic: purchases are driven by portfolio rebalancing and de-dollarisation objectives, not price momentum. This buying is broadly insensitive to short-term price levels, demand persisted as gold moved higher.
In 2025, purchases moderated to 863 tonnes (by World Gold Council), below the prior three-year pace but still above the historical average. Poland led disclosed buying; a significant share of accumulation remained unreported across multiple jurisdictions. The sovereign floor explains the otherwise anomalous divergence between rising prices and flat or declining ETF holdings through 2022–2024: the marginal buyer was sovereign, not market-driven capital.
Phase 2: Western capital has returned but remains structurally under-allocated (2025–)
The second phase began in 2025 with the re-engagement of institutional and retail flows. ETF holdings increased by approximately 801 tonnes globally (World Gold Council); total gold demand exceeded 5,000 tonnes for the year. Bar and coin demand reached multi-year highs across multiple regions. The rally transitioned from narrow to broad-based, sovereign accumulation continued while private capital added an incremental and cyclically sensitive layer.
Global Gold ETFs have been in outflow for the period 2022-2024
Source: https://www.gold.org/goldhub/data/gold-etfs-holdings-and-flows
Gold ETFs constitute approximately 0.17% of U.S. private financial assets, a low allocation that persists despite rising gold prices. Historical norms for gold within diversified institutional portfolios typically range from 1–2% of assets under management. A reversion toward the lower end of that range, without any change in central-bank demand, implies several hundred additional tonnes of ETF inflows per year. Phase 2 has room to expand from within its existing buyer set.
Phase 3: Crypto-native demand is integrating gold as productive collateral
A third demand layer has emerged within crypto-native financial infrastructure. In absolute terms it remains small relative to Phase 1 and Phase 2. In structural terms it introduces mechanisms that did not exist in prior gold cycles, and its price sensitivity profile differs from both central banks and ETF investors.
Tether’s USDT Reserve Assets
Stablecoin reserve accumulation represents the largest near-term buyer within this channel. Tether’s USDT is backed by a diversified reserve pool of approximately $190B, the majority held in US Treasuries and money market instruments (74%), but with a structurally significant allocation to gold (10%) and Bitcoin (3.5%). This reserve composition is a deliberate policy choice. It places Tether outside the framework of US stablecoin legislation: the GENIUS Act, which passed the US Senate in May 2026, requires compliant issuers to back stablecoins exclusively with USD cash, short-dated Treasuries, or Fed reserves, explicitly excluding gold, Bitcoin, and non-USD assets. Tether, incorporated offshore and not subject to US jurisdiction, operates under no such constraint.
Tether USDT’s Reserve
The distinction matters structurally. Compliant US stablecoins will direct reserve growth entirely into short-duration dollar instruments. Tether’s reserve growth, driven by its own liability expansion, feeds directly into physical gold markets. Reported 2025 purchases placed Tether among the top institutional gold buyers globally, exceeded by only a small number of central banks including Poland. This demand is balance-sheet driven: it responds to USDT circulation growth, not to macro rates or portfolio rebalancing cycles.
Tokenized Gold
Tokenized gold, digital tokens backed 1:1 by physically allocated gold held in audited vaults, has grown from a niche instrument to a $6B+ market with approximately 35–40 tonnes outstanding as of early 2026. The two dominant products, Paxos Gold (PAXG) and Tether Gold (XAUT), account for the majority of supply, though a broader set of issuers has emerged on Ethereum and other settlement layers.
While the total size remains small at less than 40 tonnes, the growth rate cannot be neglected. In the past half year, the total supply of tokenized gold has doubled in quantity.
The structural significance of tokenized gold is not its current scale but its functional role: it converts a traditionally settlement-inefficient asset into programmable collateral that can be posted in DeFi lending protocols, used as margin in on-chain derivatives, and transferred across borders without the custody friction of physical gold.
Yield-bearing Tokenized Gold Product
Yield-bearing gold structures address gold’s traditional limitation as a non-yielding asset. Products such as thUSD/thGOLD allocate capital into tokenized gold, hedge price exposure via short futures, and generate yield from two sources: the futures roll in contango (where futures prices exceed spot, producing a positive roll return as contracts converge toward expiry) and lending of tokenized gold positions. This model converts gold from a passive store of value into productive collateral, capturing value from the depth of existing gold derivatives markets and redistributing it to holders.
However, these strategies are novel and carry risks that distinguish them from traditional gold exposure. The frequency of DeFi exploits in 2025–2026 has kept adoption cautious, and these risk factors constrain the pace at which yield-bearing gold products scale beyond crypto-native participants.
What the three-phase demand base implies for gold allocation
The forward implication of this sequencing is directional rather than precise. Each layer has a different growth ceiling and a different sensitivity to macro conditions.
Demand layer profile: scale vs. price sensitivity vs. growth trajectory
Central-bank demand is likely to moderate from its 2022–2024 pace at sustained elevated prices, the marginal cost of reserve diversification rises as gold’s share of total reserves increases. ETF and private demand has the largest near-term expansion potential given the under-allocation gap: if Western institutional allocations moved from 0.17% toward a conservative 0.5% of U.S. private financial assets, the implied incremental demand is in the range of 1,500–2,000 tonnes. Crypto-native demand is the smallest in absolute terms but carries the highest growth rate from a low base; its ceiling is less defined because the use cases, collateral, yield generation, reserve backing, are structurally new.
Disclaimer: The information provided herein does not constitute investment advice, financial advice, trading advice, or any other sort of advice, and should not be treated as such. All content set out below is for informational purposes only.
The post BloFin Research: Gold’s Three-Phase Demand Expansion appeared first on BeInCrypto.
