PAX Gold is issued by Paxos Trust Company, a New York limited purpose trust company chartered and supervised by the New York State Department of Financial Services (NYDFS). Each PAXG token represents one fine troy ounce of a London Bullion Market Association (LBMA) Good Delivery gold bar, allocated and held in Brink’s vaults in London (Paxos, PAX Gold product page). The token is an ERC-20 asset on Ethereum, with contract address 0x45804880De22913dAFE09f4980848ECE6EcbAf78 (Etherscan, PAXG contract), and it launched in September 2019 (PR Newswire, 17 Sep 2019). Paxos publishes monthly attestations of gold holdings; the independent accounting firm of record transitioned from WithumSmith+Brown to KPMG effective 28 February 2025 (Gate.com, PAXG regulatory overview), with the full attestation archive available on Paxos’ transparency page (Paxos, PAX Gold Transparency). Vault custodians named in Paxos’ disclosures include Brink’s alongside other LBMA-approved custodians such as HSBC and ICBC Standard Bank (LlamaRisk, PAXG risk assessment). PAXG’s market capitalization has fluctuated with the gold price and net token issuance, ranging roughly between $1.8 billion and $2.6 billion across various points in 2026, with the token itself touching an all-time high near $5,622 in January 2026 as spot gold rallied (CoinMarketCap, PAX Gold, 2 Jul 2026; OpenPR, PAXG price analysis). By February 2026, market trackers reported the combined tokenized-gold sector — led by PAXG and XAUT — surpassing $6 billion in aggregate market capitalization (AInvest, 22 Feb 2026).
Tether Gold (XAUT) is issued by TG Commodities Limited, an entity affiliated with Tether and registered in the British Virgin Islands, with token administration tied to Tether’s broader corporate structure historically anchored in Malta and El Salvador. Each XAUT represents one fine troy ounce of physical gold held in Swiss vaults to the LBMA Good Delivery standard, and the token is issued natively on both Ethereum (ERC-20) and Tron (TRC-20) (Binance Academy, What Is Tether Gold). Tether’s Q4 2025 attestation confirmed 520,089.35 troy ounces of gold backing XAUT in circulation, with roughly 409,217 ounces already sold and the remainder available for future issuance (MEXC News, Tether Gold Q4 2025 attestation). By late 2025 Tether disclosed holding on the order of 140 tonnes of physical gold across its balance sheet and the XAUT reserve pool, worth roughly $24 billion at prevailing prices — a scale Bloomberg described as the largest known hoard outside sovereign, ETF, and commercial-bank vaults (MEXC News, citing Bloomberg reporting on Tether gold holdings). XAUT’s market capitalization has run consistently ahead of PAXG’s through 2026, with various trackers placing it between $2.4 billion and $3.4 billion (CoinStats, Tether Gold investment analysis; BaltEX, XAUT review 2026), and Tether has stated XAUT represents roughly 60% of the total gold-backed stablecoin market by value (Tether.io, XAUT market share announcement).
Kinesis KAU departs from the one-ounce convention used by PAXG and XAUT: one KAU represents one gram of allocated gold of at least 9999 fineness. The issuing structure runs through Kinesis Money, with vaulting and minting governed by KMS Labs entities and a Cayman Islands trust arrangement, and custody handled by Loomis International together with a network of Allocated Bullion Exchange (ABX) partner vaults spanning Zurich, London, New York, Singapore, Hong Kong, Dubai, Istanbul, Vaduz, Toronto, Panama City, Batam, Brisbane, and Sydney (Kinesis Money, Gold KAU Minting Programme White Paper v1.1, May 2025). Kinesis publishes a “trust and security” overview describing biannual physical audits conducted by independent inspection firms such as Bureau Veritas and Inspectorate International (Kinesis Money, Trust & Security Detailed Overview). KAU exists both on the purpose-built Kinesis Blockchain (a Cosmos SDK-based chain with its own Kinesis Monetary System) and as a bridged ERC-20 representation on Ethereum, giving holders a 1:1 claim on the Cayman-held physical gold regardless of which chain the token sits on. Kinesis differentiates itself commercially by paying a yield to holders and traders drawn from a transaction-fee pool, a feature the pure custodial-claim tokens (PAXG, XAUT) do not offer.
PGOLD, branded as “Pleasing Golden,” launched in October 2025 from Pleasing International, with each PGOLD token representing one troy ounce of LBMA-certified gold held in Hong Kong vaults and priced via a Chainlink-powered spot feed; the issuer has signaled plans to expand vaulting into other APAC hubs and Dubai, and has deployed the token on Arbitrum One and ApeChain rather than restricting it to a single chain (GlobeNewswire, 29 Oct 2025). Notably, Hong Kong’s Stablecoins Ordinance explicitly excludes gold-referenced tokens from its licensing scope, meaning PGOLD and similar products currently sit outside that specific regime even as the city positions itself as a gold-trading hub (AInvest, Hong Kong stablecoin and gold integration analysis). HashKey Exchange, a Hong Kong Securities and Futures Commission (SFC) Type 1 and Type 7 licensee since November 2022, began offering OTC trading of Tether Gold (XAUT) in January 2026 and separately launched a tokenized, non-listed “T-Class” product tracking the Hang Seng Gold ETF (HSGLD) in April 2026 (TechFlow, HashKey XAUT OTC launch; Binance Square, HashKey tokenized Hang Seng Gold ETF; Bloomberg, HashKey licensing, 3 Aug 2023). Separately, HSBC launched a retail-facing “HSBC Gold Token” for its Hong Kong customers in March 2024, letting retail banking clients buy fractional, tokenized exposure to physical gold through the bank’s own app (The Digital Banker, HSBC Gold Token launch).
Backed Finance, a Swiss-regulated issuer, offers bGOLD as an allocated, Swiss-vaulted gold token structured to align with the European Union’s Markets in Crypto-Assets Regulation (MiCA) disclosure and reserve conventions, though — as detailed in Section 6 below — no gold-referencing token had actually completed full MiCA asset-referenced-token authorization as of early 2026. Backed’s tokens follow an authorized-participant primary-market model similar to exchange-traded commodity (ETC) structures, where large institutional participants create and redeem tokens directly against physical gold, while retail investors transact in the secondary market on-chain.
Digix (DGX), one of the earliest gold-tokenization projects, represented one gram of gold held in Singapore and Canadian vaults; the DigixDAO governance structure wound down activity by 2020 as trading volumes and redemption demand collapsed, illustrating that a working custody chain alone does not guarantee commercial viability without sustained liquidity and market making. OneGram (OGC), marketed from Dubai with Sharia-compliance positioning, similarly failed to sustain meaningful secondary-market liquidity after its initial coin offering phase. AABBG, a gold-backed token associated with an Asia-based exchange group, saw its trading and redemption infrastructure fall into disuse as the sponsoring platform’s broader business contracted. The common failure pattern across dormant gold tokens is not fraud or missing metal so much as an inability to sustain two-sided market-maker liquidity once initial promotional trading interest faded — a risk every current entrant, including PGOLD and HashKey’s newer products, still has to prove it can avoid at scale.
Both PAXG and XAUT trade across a broad set of centralized exchanges — Binance, Kraken, Coinbase, Bitfinex, Bybit, OKX, and Gate among others — with XAUT typically posting higher 24-hour spot volume, driven in part by Tether’s deep pre-existing liquidity relationships and Bitfinex’s role as its single largest spot venue (Perfolio, XAUT vs PAXG comparison, 19 May 2026). Aggregate monthly spot trading volume for the two tokens averaged roughly $5.7 billion (PAXG) and $5.3 billion (XAUT) over a trailing fifteen-month period reported in mid-2026, with combined tokenized-gold spot turnover for the first quarter of 2026 reported at roughly $90.7 billion (Binance Square, Tokenized Gold Q1 2026 Trading Volume). Derivatives markets have grown even faster: aggregate open interest across major venues reached roughly $293 million for XAUT and $205 million for PAXG by early April 2026, with XAUT perpetual daily volume surging from under $100 million in late 2025 to peaks above $900 million in March 2026 following Binance’s December 2025 XAUT perpetual listing (Ethena Governance Forum, Expand USDe Backing with Tokenized Gold, 16 Apr 2026). Retail-facing platforms including WEEX also list leveraged PAXG-USDT and XAUT-USDT futures contracts, occasionally running zero-fee promotions on both tokens to attract volume (WEEX, Top Exchanges Listing PAXG and XAUT in 2026).
Neither PAXG nor XAUT charges an ongoing storage or custody fee, a deliberate design choice that both issuers absorb as a cost of doing business rather than passing on to passive holders, treating it as a loss leader against issuance and redemption revenue. Paxos charges a tiered creation and redemption fee when converting between USD, unallocated gold, and PAXG directly on Paxos.com — for example, transactions between 2 and 25 PAXG have historically incurred a 1.00% fee, with lower rates at larger institutional sizes — plus a small 0.02% fee embedded in on-chain PAXG transfers themselves, alongside standard Ethereum gas costs (Flashift, PAXG vs XAUT fee comparison, 3 Jul 2026). Tether charges a flat 0.25% one-time fee on direct issuance and redemption of XAUT, with no separate embedded on-chain transfer fee beyond standard blockchain gas costs, meaning frequent traders moving XAUT wallet-to-wallet on Tron in particular can pay near-zero marginal cost per transfer (Paxos — PAX Gold Official). Minimum direct-purchase sizes also differ meaningfully: Paxos.com allows purchases as small as roughly 0.03 PAXG, while Tether Gold’s direct minimum purchase is 50 XAUt tokens, equivalent to 50 troy ounces of gold — though both tokens trade in far smaller increments on secondary exchanges (Flashift, PAXG vs XAUT minimum purchase comparison).
Ethereum hosts the original and largest-cap deployments of both PAXG and XAUT as ERC-20 tokens, along with Backed’s bGOLD and most of the smaller entrants such as PGOLD (on Arbitrum, an Ethereum layer-2). PAXG’s Ethereum contract (Etherscan, PAXG token contract) and XAUT’s Ethereum contract (DeFi Terminal, Tether Gold asset page) are the addresses referenced by essentially every exchange listing, custody attestation, and DeFi integration for these tokens. Ethereum’s dominance reflects its deep composability with lending markets (Aave), decentralized exchanges (Curve, Uniswap), and perpetuals platforms (GMX) that gold-token holders use to borrow against or trade their positions without selling the underlying claim — a use case explored further in Section 8.
Tether issues XAUT natively on Tron (TRC-20) in addition to Ethereum, a design choice consistent with Tether’s broader USDT strategy of using Tron for lower transaction fees and faster settlement in Asia-facing retail and remittance corridors (Binance Academy, What Is Tether Gold). Because Tron transaction costs run a small fraction of typical Ethereum gas fees, exchanges and OTC desks serving retail users in Southeast Asia and elsewhere often default to the Tron deployment for smaller transfers, while institutional and DeFi-focused holders gravitate toward the Ethereum version for composability with lending and trading protocols.
Solana hosts wrapped or bridged representations of gold tokens rather than native issuance by Paxos or Tether directly, typically created by third-party bridge protocols that lock the Ethereum-native token and mint a corresponding Solana SPL token, inheriting both Solana’s faster/cheaper transaction environment and the additional smart-contract risk of the bridging mechanism itself. XRPL (the XRP Ledger) and TON (The Open Network) have both seen gold-token or gold-referencing DeFi products emerge from third-party issuers seeking to tap XRPL’s payment-settlement user base and TON’s Telegram-integrated distribution, though none of these approaches the market capitalization of the Ethereum-native PAXG or XAUT deployments. As with Solana, any non-native deployment on these chains depends on a bridge or wrap mechanism whose security is distinct from — and sometimes weaker than — the base token’s own smart contract and custodial arrangement.
Kinesis operates its own Kinesis Blockchain, a Cosmos SDK-derived network running the Kinesis Monetary System (KMS), designed specifically to support KAU (gold) and KAG (silver) as native monetary units rather than as one asset among many on a general-purpose smart-contract platform (Kinesis Money, Trust & Security Detailed Overview). This purpose-built design lets Kinesis embed yield distribution and transaction-fee mechanics directly into the base chain rather than relying on external DeFi protocols, but it also means KAU liquidity is concentrated within the Kinesis ecosystem and its exchange partners rather than spread across the broader multi-chain DeFi landscape that PAXG and XAUT participate in.
Every bridged deployment of a gold token — moving a claim from its native chain to Solana, for instance — depends on bridge infrastructure that has proven to be one of the most exploited categories in all of crypto. The Wormhole bridge connecting Solana and Ethereum suffered a hack in February 2022 resulting in the loss of roughly 120,000 wrapped ETH (worth around $325 million at the time), later covered by its parent company Jump Crypto to preserve peg integrity. The Nomad bridge was exploited in August 2022 for roughly $190 million due to a faulty contract initialization that allowed near-arbitrary message replay, and the Ronin bridge supporting the Axie Infinity ecosystem was drained of roughly $625 million in March 2022 after attackers compromised validator private keys. None of these three incidents directly targeted gold tokens, but they illustrate the structural risk that any bridged gold-token representation on a non-native chain carries an additional attack surface layered on top of the underlying custodial and audit risk of the physical gold itself — a risk that does not exist for a token used only on its issuer’s native chain.
Tether has pursued the most aggressive multi-chain strategy among major gold-token issuers, extending XAUT beyond its original Ethereum and Tron deployments; third-party trackers documented a distinct “XAUt0” full-chain deployment product generating roughly $100 million in average monthly trading volume alongside the primary XAUT token, illustrating how issuers now treat chain footprint itself as a liquidity and distribution lever rather than a one-time technical decision (Binance Square, Tokenized Gold Q1 2026 Trading Volume analysis). Paxos, by contrast, has kept PAXG concentrated on Ethereum, arguing that deep composability with Ethereum-native DeFi protocols and regulated exchange integrations outweighs the lower transaction costs available on faster chains — a strategic trade-off that shows up directly in the derivatives and DeFi collateral data discussed in Section 8, where PAXG dominates Ethereum-based lending markets while XAUT captures a larger share of high-frequency, low-value transfer volume on Tron.
Chain selection also carries regulatory weight: a token issued natively only on a single, well-audited chain gives regulators and auditors a single, unambiguous ledger of record, whereas multi-chain and bridged deployments create parallel, and sometimes reconciled with meaningful lag, supply figures that attestation providers must aggregate across networks before publishing a single combined reserve figure. This is one reason Paxos’ monthly PAXG attestations (Paxos, PAX Gold Transparency) remain comparatively simple to audit relative to a hypothetical gold token spread natively across five or six chains, and it is a structural consideration that newer entrants such as PGOLD — already deployed on both Arbitrum One and ApeChain at launch — will need to manage as their attestation and audit practices mature.
Ethereum gas costs remain a real, if often underappreciated, friction point for gold-token holders: typical PAXG transfers can cost anywhere from a few dollars to $20 or more during normal network conditions, and spike to $50 or higher during periods of network congestion driven by unrelated activity such as NFT mints or memecoin trading surges. XAUT transfers on Tron, by contrast, typically cost fractions of a cent and settle in seconds, which is precisely why high-frequency traders moving gold exposure between exchanges multiple times per day gravitate toward the Tron deployment even though the Ethereum deployment offers deeper DeFi composability. This cost asymmetry is a structural, not incidental, feature of the two products’ chain strategies, and it materially shapes which token dominates which use case — passive DeFi collateral (PAXG on Ethereum) versus active, high-turnover trading (XAUT on Tron).
The LBMA Gold Price is administered by ICE Benchmark Administration (IBA) under UK Financial Conduct Authority regulation, running a twice-daily electronic auction at 10:30 and 15:00 London time across USD, EUR, and GBP (ICE Developer Portal, LBMA Gold Price). This auction — the direct successor to the London Gold Fixing that ran continuously since 1919 — remains the reference point that essentially every gold-backed token issuer cites when describing how it values its physical reserves, even though none of the tokenized-gold products settle directly against the auction itself (ICE, 2 Feb 2015).
CF Benchmarks, a UK FCA-authorized benchmark administrator operating under the UK Benchmarks Regulation (BMR), publishes a family of real-time reference rates for crypto-native assets, and its methodology framework — built around continuous, volume-weighted aggregation across constituent exchanges — is the same approach the firm applies to its crypto asset indices used by regulated products such as exchange-traded funds. For gold-linked tokens, having a BMR-regulated administrator in the reference-price chain matters directly to the regulatory classification questions discussed in Section 6, since UK and EU benchmark rules impose specific governance, oversight-committee, and conflict-of-interest requirements on anyone administering a rate used in a regulated financial product.
Chainlink operates decentralized XAU/USD price feeds aggregating data from multiple independent node operators and data providers, and its feeds underpin the price logic for several gold-tokenization platforms, including the Chainlink-powered spot market referenced by Pleasing International’s PGOLD launch (GlobeNewswire, PGOLD launch press release). Chainlink’s decentralized oracle model — multiple independent nodes submitting price observations that are aggregated on-chain — is designed to resist single-point manipulation, a critical property for any protocol that uses a gold price feed to determine loan-to-value ratios or liquidation thresholds for gold-token collateral on lending markets such as Aave.
Pyth Network takes a different architectural approach from Chainlink, sourcing price data directly from first-party publishers — trading firms, exchanges, and market makers — who push their own proprietary price observations on-chain, which Pyth then aggregates into a confidence-weighted price and uncertainty band. Pyth’s XAU/USD feed is widely used across Solana-native and multi-chain DeFi protocols that need low-latency gold pricing, and its confidence-interval design gives protocols an explicit signal of price uncertainty rather than a single point estimate, which is particularly useful during periods of thin gold-market liquidity such as major holiday windows in key trading centers.
Chronicle emerged from the oracle infrastructure originally built for MakerDAO (now Sky), and it continues to provide price feeds — including precious-metals references — to DeFi protocols that value MakerDAO’s original security-first, multi-signer “Oracle Security Module” design lineage, where price updates pass through a time-delayed queue specifically to blunt the impact of a compromised or manipulated single feed. RedStone offers a more modular “on-demand” oracle model, including a dedicated Metals Price API, that pushes price data to smart contracts only when a transaction needs it rather than continuously posting updates on-chain, which can reduce gas costs for protocols that need gold pricing only intermittently rather than in every block.
The Shanghai Gold Exchange (SGE) Benchmark Price, launched 19 April 2016, runs a twice-daily, centrally cleared auction at 10:15 and 14:15 Beijing time on physically delivered 1 kg bars of 99.99% purity gold, and it frequently trades at a premium or discount to London and COMEX levels reflecting Chinese domestic supply-demand and capital-control dynamics rather than arbitrage failure (BullionStar, SGE infrastructure guide). While no major gold token currently references SGE as its primary settlement benchmark, the SGE/London basis remains an important signal that Asia-facing tokenized-gold platforms — including HashKey’s Hong Kong-listed products — monitor closely given the region’s growing role in both physical gold trading infrastructure and digital-asset distribution.
The practical stakes of oracle selection become clearest inside lending protocols. When Aave governance evaluated adding XAUT as collateral on its v3 Core Instance, the proposal explicitly weighed oracle reliability and market depth alongside aggregate open interest and exchange liquidity thresholds before approving the listing in mid-2025 and later adjusting supply caps as usage grew (Aave Governance Forum, ARFC: Add XAUt to Aave v3 Core Instance, 19 Jun 2025). A stale, manipulated, or wide-confidence-interval price feed directly threatens the protocol’s solvency: if the oracle reports a gold price meaningfully above the true market price, borrowers can extract more stablecoin debt than their collateral actually supports, while a price reported below market can trigger unnecessary liquidations of otherwise healthy positions. This is precisely why protocols increasingly combine multiple oracle sources — for instance, cross-checking Chainlink against Pyth’s confidence-weighted feed — rather than relying on a single provider for a collateral type as economically significant as gold-backed tokens have become.
Because the LBMA Gold Price, CF Benchmarks’ rates, and the underlying oracle feeds are all published with defined methodologies and update schedules, they double as a structured, machine-readable reference layer that AI systems, index providers, and compliance tools increasingly query directly rather than relying on scraped spot-price pages. IBA’s decision to consolidate administration of gold, silver, platinum, and palladium benchmarks under a single infrastructure provider from 1 July 2026 is itself a step toward exactly this kind of standardized, single-source reference architecture across the precious-metals complex (Business Wire, IBA consolidates LBMA precious-metals benchmarks, 7 Jul 2026).
A subtle but important distinction in the gold-token pricing stack is that the price an oracle reports for XAU/USD is not the same data point as the net asset value an issuer like Paxos or Tether uses internally to manage creation and redemption of new tokens. In practice, secondary-market trading prices for PAXG and XAUT track the oracle-fed spot price closely because arbitrageurs can create or redeem tokens directly with the issuer whenever a persistent premium or discount opens up on exchanges, buying low from one venue and redeeming or minting at the issuer’s reference price to capture the spread. This arbitrage mechanism is what keeps PAXG and XAUT trading within a narrow band of the London spot fix on liquid venues, and it depends on the oracle layer, the issuer’s own reference pricing, and exchange order books all remaining loosely synchronized — a three-way check that has generally held up well through 2025 and 2026 even during periods of sharp gold-price volatility.
Paxos holds the physical gold backing PAXG in Brink’s vaults in London, alongside disclosed relationships with other LBMA-approved custodians including HSBC and ICBC Standard Bank (LlamaRisk, PAXG risk assessment). Paxos publishes monthly independent attestations verifying that token supply is fully backed by allocated LBMA Good Delivery bars, with KPMG now serving as the attesting firm after taking over from WithumSmith+Brown at the end of February 2025 (Gate.com, PAXG regulatory overview), with the reports themselves published on Paxos’ transparency page (Paxos, PAX Gold Transparency).
Tether Gold’s reserves and Backed’s bGOLD both rely on Switzerland’s refiner-vault ecosystem, the same cluster of facilities that anchors much of Europe’s LBMA Good Delivery gold supply: Argor-Heraeus (Chiasso and Mendrisio), Metalor Technologies (Neuchâtel), and Valcambi (Balerna). These refiners combine on-site vaulting, LBMA Good Delivery accreditation, and refining capacity in a single facility, which lets a gold-token issuer source, refine, cast, and store bars without moving metal between third parties — a meaningful chain-of-custody simplification compared with issuers who must contract vaulting and refining separately. Tether’s Q4 2025 attestation of 520,089.35 fine troy ounces of gold backing XAUT is published against this Swiss custodial base (MEXC News, Tether Gold Q4 2025 attestation).
Kinesis anchors KAU custody through Loomis International together with its Allocated Bullion Exchange (ABX) partner-vault network spanning more than a dozen cities globally — including Zurich, London, New York, Singapore, Hong Kong, Dubai, Istanbul, Vaduz, Toronto, Panama City, Batam, Brisbane, and Sydney (Kinesis Money, Gold KAU Minting Programme White Paper v1.1). Kinesis discloses that physical stock is subject to biannual independent audits by inspection firms such as Bureau Veritas and Inspectorate International (Kinesis Money, Trust & Security Detailed Overview), a lower frequency than Paxos’ monthly cadence but broader in geographic distribution than the single-city concentration of PAXG’s Brink’s London holdings.
Hong Kong-facing products, including HashKey’s tokenized gold offerings and PGOLD, rely on vaulting standards associated with the Hong Kong Bullion Association and a physical infrastructure buildout that includes the Hong Kong International Airport (HKIA) gold vault, which authorities have targeted for expansion from roughly 150 tonnes of capacity toward approximately 1,000 tonnes in the near term and more than 2,000 tonnes by 2028, alongside the Shanghai Gold Exchange’s first offshore delivery vault, opened in Hong Kong in June 2025 (FinancialContent, Hong Kong commodity hub buildout, 12 Nov 2025; Golden Ark Reserve, Hong Kong gold market infrastructure).
Audit frequency varies meaningfully by issuer: Paxos publishes monthly attestations for PAXG, Kinesis discloses biannual physical audits for KAU, and most LBMA-member vault operators undergo their own quarterly or annual Good Delivery List compliance reviews independent of any specific token issuer’s reporting cycle. Beyond attestation frequency, the underlying LBMA Good Delivery framework itself requires accredited refiners to maintain minimum tangible net worth, proven refining history, and ongoing production-quality performance to remain on the List, with the LBMA able to suspend or remove refiners that fail ongoing monitoring. Vault operators in this ecosystem typically carry commercial insurance placed through the Lloyd’s of London market, covering physical loss, theft, and certain custodial risks, though the specific policy limits and named insurers for any individual token issuer’s vault arrangement are rarely disclosed publicly at the line-item level.
Physical bullion held in LBMA-accredited vaults is conventionally insured through the Lloyd’s of London market, where specialist syndicates underwrite precious-metals risk as part of a broader fine-art-and-specie insurance category covering theft, physical loss, and certain transit risks (HWS Specialty, Insurance for Gold, Bullion and Precious Metals). Lloyd’s has separately extended into digital-asset-adjacent insurance products, including a cryptocurrency wallet insurance solution for custody providers, reflecting the market’s increasing willingness to underwrite hybrid physical-and-digital custodial risk of exactly the kind that gold-token issuers carry (Lloyd’s, Cryptocurrency Wallet Insurance Solution launch). That said, individual token issuers rarely publish the specific policy limits, named syndicates, or per-vault coverage caps backing their reserves, leaving institutional counterparties to rely largely on the reputational and regulatory standing of the custodian (Brink’s, Loomis) and the attesting accountant rather than on any publicly verifiable insurance certificate.
Paxos’ status as an NYDFS-chartered limited-purpose trust company subjects PAX Gold to a materially different supervisory regime than most competing products: NYDFS imposes ongoing capital, custody, and compliance requirements directly on Paxos as a regulated entity, independent of whatever attestation cadence Paxos chooses to publish voluntarily. Some competing attestation regimes run on a slower cadence — for instance, certain gold-token issuers report quarterly third-party attestations rather than Paxos’ monthly schedule — which matters directly to institutional counterparties evaluating counterparty and reserve risk, since a longer gap between attestations gives less frequent visibility into potential supply/reserve mismatches.
A meaningful custodial distinction across issuers is whether a token holder’s claim traces to a specific, serial-numbered bar or to a pooled position within the vault’s aggregate holdings. Paxos markets PAXG as tracing to specific, identifiable bars within its Brink’s London holdings, with bar-level detail available as part of its transparency disclosures, while some competing products describe custody in terms of pooled, allocated gold without necessarily publishing bar-by-bar serial number detail to every holder. Neither approach is inherently fraudulent — both can satisfy the LBMA’s allocated-gold definition — but the serial-number-level disclosure model gives institutional counterparties and auditors a finer-grained verification path when reconciling token supply against the vault’s physical bar list.
The London Bullion Market Association Good Delivery standard defines bars weighing between 350 and 430 troy ounces (roughly 12.4–13.4 kg) with a minimum fineness of 995 parts per thousand (99.5% pure gold), cast or produced by one of roughly 70 LBMA-accredited refiners worldwide that meet the Association’s ongoing standards for refining capability, financial strength, and bar-marking accuracy. This is the bar standard underlying essentially every major tokenized-gold reserve, including PAXG’s Brink’s London holdings and XAUT’s Swiss-refiner reserves, and it is also the settlement standard referenced by the twice-daily LBMA Gold Price auction itself (Bullion Trading LLC, LBMA/COMEX pricing explainer).
COMEX gold futures, operated by CME Group, specify a standard contract size of 100 troy ounces with a minimum fineness of 995, smaller than the LBMA’s 400-ounce convention, in addition to smaller e-micro and enhanced-delivery contract variants (CME Group, Gold Futures Contract Specs). CME’s Accumulated Certificates of Exchange (ACE) mechanism allows a standard 400-ounce Good Delivery bar to be split into four tradeable 100-ounce units eligible for COMEX delivery, providing a structural bridge between the London 400-oz wholesale standard and COMEX’s smaller contract denomination (CME Group, Gold (Enhanced Delivery) Futures FAQ). No major tokenized-gold issuer currently sources reserves directly against COMEX warehouse stock, but the ACE mechanism illustrates the general principle — splitting large wholesale bars into smaller certified units — that underlies how token issuers translate 400-oz LGD holdings into one-ounce or one-gram token denominations at the accounting level.
The 1,000-gram kilobar (roughly 32.15 troy ounces), typically cast to 999 or 9999 fineness, is the dominant physical trading and delivery unit across Asian gold markets, including the Shanghai Gold Exchange’s benchmark settlement lots and much of Hong Kong, Singapore, and Indian wholesale trade (BullionStar, SGE infrastructure guide). Kinesis’ gram-denominated KAU token is conceptually closer to the kilobar tradition than to the troy-ounce LGD convention, and Hong Kong-facing tokenized products are more likely to reference kilobar-based physical delivery infrastructure as the region’s vault capacity expands.
At the smallest end of the physical market sit retail bars and coins — one-ounce products such as the Krugerrand, American Gold Eagle, or Canadian Maple Leaf, as well as multi-denomination products like PAMP Suisse’s CombiBar, which is scored so that individual one-gram segments can be snapped off and sold separately. These retail formats are the physical-market analogue to what tokenization achieves digitally: taking a large, illiquid wholesale unit and dividing it into pieces small enough for individual investors to hold, though retail bars carry meaningfully higher fabrication premiums per ounce than either LGD wholesale bars or a tokenized gram-denominated claim.
Every major tokenized-gold issuer holds its physical reserves as LGD bars while issuing tokens in units far smaller than a single bar — PAXG and XAUT at one troy ounce (meaning roughly 350–430 tokens correspond to a single physical bar), and KAU at one gram (meaning roughly 11,000–13,400 tokens correspond to a single bar, since one troy ounce equals approximately 31.1 grams). This is purely an internal accounting and custody-ledger exercise for the issuer: the physical bar is never actually cut or divided, and the issuer’s attestation confirms that total token supply in circulation is matched, at the aggregate level, by total fine-ounce holdings recorded in the vault’s bar list. This is also precisely why large-scale physical redemption — discussed in Section 7 — typically requires redeeming a minimum quantity of tokens equivalent to at least one full LGD bar, rather than allowing redemption of an arbitrary fractional amount into physical form.
Where a physical LGD bar cannot practically be divided below roughly a one-gram fabricated unit (the smallest widely available retail format, such as individual segments of a PAMP CombiBar), tokenized products push fractional ownership far further at the accounting layer. PAXG is divisible to 18 decimal places — the standard precision of an ERC-20 token on Ethereum — letting a holder own an economically meaningless but technically valid fraction of a single fine troy ounce, while XAUT is divisible to six decimal places (Flashift, PAXG vs XAUT divisibility comparison). This divisibility is purely a ledger-entry construct: the underlying LGD bar is never physically altered, and Paxos’ and Tether’s attestations reconcile only at the aggregate, whole-bar level of the vault’s bar list, not against any individual fractional token holding.
The credibility of every bar-size standard discussed above — LGD, COMEX, kilobar, or retail — ultimately depends on the accrediting body’s ongoing monitoring of the refineries that cast and stamp the bars. The LBMA’s Good Delivery List requires accredited refiners to maintain minimum tangible net worth thresholds, a demonstrated multi-year refining track record, and passing performance in periodic, unannounced proficiency testing of assay accuracy; refiners that fail ongoing monitoring can be suspended or removed from the List entirely. Because every major tokenized-gold issuer sources exclusively from LGD-accredited refiners — Argor-Heraeus, Metalor, and Valcambi for the Swiss-vaulted products, and the Brink’s-network refiners for PAXG — this refiner-level accreditation discipline is arguably a more important quality backstop for token holders than any blockchain-level verification mechanism, since no smart contract can independently confirm the fineness or weight of a physical bar sitting in a vault.
No blockchain mechanism can independently verify that a physical bar sitting in a Brink’s or Loomis vault actually meets its stated fineness — that verification happens entirely through the traditional physical-assay and refiner-accreditation process described above, with the blockchain layer only recording and transferring claims against gold whose physical properties have already been certified off-chain. This is a fundamental and often underappreciated limit of gold tokenization as a category: the “trust-minimization” properties that make blockchain assets attractive in purely digital contexts (verifiable scarcity, transparent transaction history) do not extend to the physical-world assay and custody chain that ultimately backs the token, which instead depends on the same refiner accreditation, vault audit, and attestation infrastructure that has underpinned physical bullion markets for decades.
Under the EU’s Markets in Crypto-Assets Regulation (MiCA), tokens that purport to maintain stable value by referencing a commodity, a basket of currencies, or a basket of crypto-assets — rather than a single official currency — are classified as Asset-Referenced Tokens (ARTs) under Titles III and IV, with the stablecoin provisions of MiCA applying from 30 June 2024 and the main crypto-asset service provider (CASP) authorization regime applying from 30 December 2024 (Chainalysis, MiCA Stablecoin Regime, 30 Jun 2024). PAX Gold and Tether Gold both qualify conceptually as ARTs because they reference physical gold rather than a fiat currency (Eco, MiCA EMTs vs ARTs Explained). Yet as of January 2026, market trackers confirmed there were still zero authorized ARTs anywhere under MiCA, despite ARTs constituting roughly a third of the regulation’s overall stablecoin framework, with the core obstacle being persistent disagreement among EU member-state regulators over whether a gold-backed token should be classified as a MiCA ART or instead as a financial instrument subject to MiFID II (MEXC, Europe Still Has Just 17 Authorized Issuers in 2026; Levi Brocklehurst, MiCA ART authorization tracker, Jan 2026). The European Securities and Markets Authority (ESMA) issued a public statement in January 2025 directing national regulators to ensure CASPs stop offering non-MiCA-compliant ARTs and EMTs for trading by the end of Q1 2025, with a “sell only” wind-down window for existing holders (ESMA, Public Statement on Stablecoins, Jan 2025). The final EU-wide MiCA transitional period for firms relying on national grandfathering ended 1 July 2026 (AI News Crypto, MiCA ART/EMT and CASP timeline).
Hong Kong’s Stablecoins Ordinance was gazetted on 30 May 2025 and became effective 1 August 2025, establishing a Hong Kong Monetary Authority (HKMA) licensing regime for fiat-referenced stablecoin issuers — but the ordinance explicitly excludes gold-referenced tokens from its scope, leaving PGOLD, HashKey’s tokenized gold offerings, and similar Hong Kong-facing products regulated (if at all) under general Securities and Futures Commission (SFC) licensing categories rather than a dedicated stablecoin framework (AInvest, Hong Kong Stablecoin Innovation and Gold Integration, Jan 2026). Hong Kong’s Financial Secretary has signaled interest in eventually extending regulatory attention to gold-linked stablecoins as the city simultaneously builds out physical gold-trading infrastructure, including the London Metal Exchange’s approval of Hong Kong as a warehouse location in January 2025 and the Shanghai Gold Exchange’s first offshore delivery vault in Hong Kong, opened June 2025 (Golden Ark Reserve, Hong Kong gold market infrastructure overview). HashKey Exchange operates under SFC Type 1 (dealing in securities) and Type 7 (providing automated trading services) licenses secured in August 2023, with a Type 9 (asset management) license layered on for certain product structures (Bloomberg, HashKey licensing, 3 Aug 2023).
The UK does not yet have a bespoke statutory regime for gold-backed tokens comparable to MiCA, but UK-facing benchmark infrastructure relevant to gold tokenization — including the LBMA Gold Price itself and CF Benchmarks’ crypto reference rates — operates under the Financial Conduct Authority (FCA)’s supervision and the UK Benchmarks Regulation (BMR), the UK’s retained and adapted version of the EU BMR framework. Any gold-token issuer wishing to reference a UK BMR-regulated benchmark, or to operate benchmark administration itself from London, must satisfy the BMR’s governance, oversight-function, and conflicts-of-interest requirements, mirroring the same structural obligations that govern the LBMA Gold Price auction administered by ICE Benchmark Administration.
Singapore’s Monetary Authority of Singapore (MAS) regulates digital payment token services under the Payment Services Act (PSA), requiring licensing for entities that deal in or facilitate exchange of digital payment tokens, a category that in practice captures much crypto-exchange and custody activity touching gold tokens even though gold-backed tokens are not singled out as a distinct regulated category. Singapore’s gold-tokenization landscape gained a significant institutional signal in mid-2026 when DBS Bank announced plans to offer tokenized gold to retail customers, marking one of the first major-bank-led retail tokenized-gold distribution efforts in Asia (Fidelity News (CoinDesk), DBS tokenized gold for retail, 11 Jun 2026).
Switzerland’s FINMA classifies crypto-assets across payment, utility, and asset (investment) token categories, and a gold-backed token structured to give holders a proprietary or claim-based interest in physical gold generally falls into FINMA’s asset-token category, subjecting issuance and distribution to securities-law-adjacent disclosure obligations depending on structure. Switzerland’s regulatory relevance to gold tokenization extends beyond token issuance itself: the country’s cluster of LBMA-accredited refiner-vaults — Argor-Heraeus, Metalor, and Valcambi — forms the physical custodial backbone for XAUT and Backed’s bGOLD, making Swiss vaulting law and refiner accreditation standards as important to the sector as FINMA’s token-classification guidance itself.
The US has no single federal framework for gold-backed tokens. Paxos operates PAXG through its New York State Department of Financial Services (NYDFS) limited-purpose trust charter, which imposes state-level custody, capital, and compliance requirements on Paxos as a trust company rather than treating PAXG itself as a federally regulated security or commodity. Separately, the Wyoming Special Purpose Depository Institution (SPDI) charter offers an alternative state-level path for digital-asset custodians, including firms handling tokenized commodities, while the SEC and CFTC continue to assess, on a case-by-case and enforcement-driven basis, whether specific gold-token structures constitute securities or commodity-linked instruments under existing federal statutes rather than under any gold-token-specific rule.
The United Arab Emirates runs two parallel virtual-asset regulatory regimes relevant to gold tokenization: Dubai’s Virtual Assets Regulatory Authority (VARA), which licenses virtual-asset service providers operating in the Dubai mainland and free zones (excluding the Dubai International Financial Centre), and the Abu Dhabi Global Market (ADGM)’s Financial Services Regulatory Authority (FSRA), which separately licenses virtual-asset activity within its own financial free zone. Both regimes have attracted gold-tokenization and broader commodity-tokenization interest given the UAE’s established physical gold-trading infrastructure in Dubai, though neither VARA nor the FSRA maintains a gold-token-specific licensing category distinct from their general virtual-asset or securities-token frameworks.
Ranking jurisdictions by regulatory stringency specifically for gold-backed tokens produces a counter-intuitive picture: the EU has the most detailed statutory category (ARTs under MiCA) but the least actual market approval activity (zero authorizations as of January 2026), while the US relies on a patchwork of state trust-company charters (NYDFS for Paxos) that impose real, enforceable capital and custody obligations without a dedicated federal gold-token statute at all. Switzerland sits in between, applying its established FINMA asset-token framework to gold tokens without needing bespoke gold-specific legislation, aided by the fact that Swiss refiner-vaults were already subject to rigorous accreditation and banking-secrecy-adjacent custodial standards well before tokenization existed as a category. Singapore and the UAE both regulate the service providers (exchanges, custodians) touching gold tokens under general digital-payment-token or virtual-asset licenses rather than creating a gold-specific product category, effectively regulating the distribution channel rather than the underlying instrument.
The absence of harmonized cross-border rules creates real friction for institutional investors seeking regulated gold-token exposure: a European asset manager cannot simply treat PAXG or XAUT as a MiCA-compliant ART today because no gold-backed ART has been authorized, a US-based institution must rely on New York trust-company law rather than a federal commodity or securities framework specific to gold tokens, and a Hong Kong-based institution operates in a market explicitly excluded from the jurisdiction’s own stablecoin licensing regime even as the same jurisdiction builds world-class physical vaulting capacity. This fragmentation is precisely why State Street’s and DBS Bank’s 2026 moves into tokenized-gold-adjacent infrastructure and distribution (detailed in Section 8) are being closely watched as bellwethers: large, systemically important institutions moving into this space despite the regulatory patchwork signal that commercial demand has begun to outpace the available legal clarity.
Paxos allows physical redemption of PAXG into an LBMA Good Delivery bar, but only above a minimum threshold close to a full bar — publicly discussed guidance points to a 430 fine troy ounce minimum, matching the top of the LGD 350–430 oz weight range, with Paxos charging a flat physical-delivery fee reported around $15,000 covering vaulting, transport, and insurance logistics to move an actual bar out of the Brink’s London vault network. Below that threshold, PAXG holders can redeem for cash value or simply sell the token on the open market, which is both faster and cheaper than triggering a physical delivery event. Full terms are published in Paxos’ PAX Gold documentation (Paxos, PAX Gold product page).
Tether Gold applies a lower minimum — commonly cited around 50 fine troy ounces — for standard redemption processed as value rather than physical delivery, while actual physical bar delivery out of the Swiss vault network requires reaching increments closer to the 430 fine ounce full-bar threshold, consistent with the LGD bar-size convention. Tether applies a redemption/transfer fee cited around 0.25% of the redeemed value, and third-party trackers note a typical settlement window of roughly T+5 business days for redemption processing (DeFi Terminal, Tether Gold asset page).
Kinesis is structurally the most granular of the major products: because KAU is denominated in one-gram units rather than one troy ounce, holders can mint and redeem in far smaller increments than PAXG or XAUT, transacting directly against the Allocated Bullion Exchange (ABX) network of partner vaults rather than requiring a full 400-oz bar’s worth of tokens before physical conversion becomes practical (Kinesis Money, Gold KAU Minting Programme White Paper v1.1). This gram-level granularity is a deliberate design choice distinguishing Kinesis from the institutional-scale redemption thresholds of PAXG and XAUT, positioning KAU closer to a retail-accessible bullion-purchasing product than a purely institutional reserve-claim instrument.
Backed structures bGOLD redemption around an authorized-participant (AP) model similar to conventional exchange-traded commodity (ETC) mechanics: large institutional participants create and redeem tokens directly against physical gold in the primary market, typically in large block sizes, while retail-scale holders transact in the secondary, on-chain market rather than redeeming physical metal directly themselves. This structure is explicitly designed to align with MiCA’s reserve-asset and redemption-rights requirements for Asset-Referenced Tokens, even though — as detailed in Section 6 — no gold-backed ART had received full MiCA authorization as of early 2026.
Hong Kong-facing tokenized gold products, including HashKey’s offerings and PGOLD, generally describe redemption in terms of local delivery against Hong Kong-based vault stock rather than cross-border physical shipment, an arrangement that becomes more operationally significant as the city’s vault capacity (including the HKIA gold vault expansion discussed in Section 4) continues to grow. A recurring structural question across nearly every issuer in this space is the distinction between unallocated gold — a general claim against a pool of metal, common in traditional bullion-bank trading — and allocated gold, where specific, serial-numbered bars are held for a specific owner. Most tokenized-gold issuers market their products as fully allocated, meaning each token corresponds to a specific claim on identified bars rather than a pooled, rehypothecatable position, but the practical fungibility between allocated and unallocated treatment at the point of large-scale institutional redemption remains an area where disclosure quality varies meaningfully across issuers.
Across every major product, the economics of redemption create a clear two-tier system: retail-size holders are, in practice, steered toward selling tokens on the secondary market (fast, cheap, typically settling within minutes) rather than requesting physical delivery (slow, expensive, and subject to minimum-quantity thresholds), while only holders with institutional-scale positions find physical redemption economically rational once flat fees and logistics costs are amortized across a large enough quantity of metal. This differential is not a hidden flaw so much as the intended design: it lets the tokens function as a liquid, exchange-tradeable proxy for gold exposure for the vast majority of holders, while preserving a genuine physical-redemption backstop for the institutional participants who need it, which is itself a key part of what keeps token prices closely tracking the underlying spot gold price through arbitrage.
To bridge the gap between institutional-scale direct redemption and genuinely small retail physical delivery, both major issuers have cultivated third-party partnerships that lower the practical redemption floor. Tether has established a strategic partnership with bullion dealer Alpha Bullion, through which retail-scale XAUT holders can redeem quantities as small as roughly one gram of physical gold rather than the 430-ounce full-bar threshold required for direct redemption from Tether itself — though this smaller-denomination path carries a meaningful premium, since the partner must melt or otherwise reallocate a large bar into small-format retail product and cover the associated shipping and insurance costs. Paxos similarly maintains institutional and retail redemption pathways of varying granularity depending on counterparty relationship and order size, with the common thread across both issuers being that smaller redemption sizes trade convenience for a materially higher effective cost per ounce.
Both PAXG and XAUT require holders to complete direct know-your-customer (KYC) verification with the issuer before minting or redeeming tokens directly — a requirement that does not apply to secondary-market trading of the tokens themselves on an exchange or DEX, where a counterparty simply buys or sells the existing token supply without touching the issuer’s primary market at all. This bifurcation matters directly to the redemption economics discussed above: a holder who acquired PAXG or XAUT purely through secondary-market trading, without ever completing issuer-level KYC, cannot redeem physical gold directly and must either onboard with the issuer retroactively or sell the token on the open market and forgo physical delivery altogether.
For the overwhelming majority of PAXG and XAUT holders, the redemption mechanics detailed above never come into play directly — instead, professional market makers and arbitrageurs use their own ability to create and redeem tokens with the issuer to keep secondary-market prices tightly aligned with spot gold, meaning a retail holder who simply sells PAXG or XAUT on Binance, Kraken, or Coinbase captures essentially the same economic outcome as a formal redemption would provide, minus a small bid-ask spread, without ever touching the issuer’s minimum-quantity thresholds or fee schedule at all. This is the same structural logic that keeps a physically backed gold ETF trading close to its net asset value: authorized participants handle the heavy-lifting creation/redemption arbitrage in bulk, while ordinary investors transact freely in the secondary market at prices those authorized participants have already helped discipline.
The combined tokenized-gold sector, led by PAXG and XAUT, surpassed roughly $6 billion in aggregate market capitalization by February 2026, with PAX Gold specifically gaining momentum through the first two months of the year (AInvest, PAX Gold Gains Momentum, 22 Feb 2026). A parallel and arguably more significant milestone came when crypto market maker Wintermute launched institutional over-the-counter trading for gold-backed tokens, naming PAXG and XAUT specifically as the two largest tokenized-gold products by market capitalization — a clear signal that professional liquidity providers now treat gold tokens as a distinct, tradeable institutional asset class rather than a niche crypto curiosity (CoinMarketCap Academy, Tokenized Gold Trading Tops $126B as Wintermute Enters Market, 17 Feb 2026).
Hong Kong pursued an aggressive, multi-track buildout of physical gold and commodity infrastructure across 2025 and into 2026: the London Metal Exchange approved Hong Kong as a warehouse location in January 2025, approved its first four LME-licensed warehouse facilities in April 2025, expanded to seven approved facilities by May 2025, and saw the first batch begin physical operations in July 2025 (London Metal Exchange, press release, 15 Apr 2025; Kitco News, LME approves three more Hong Kong warehouses, 21 May 2025; Bloomberg, LME approves Hong Kong as warehouse location, 20 Jan 2025). In parallel, the Shanghai Gold Exchange opened its first offshore gold delivery vault in Hong Kong in June 2025, extending the mainland exchange’s reach into the offshore market and building on the Shanghai-Hong Kong Gold Connect corridor active since 2015 (Golden Ark Reserve, Hong Kong gold market infrastructure). The city’s Stablecoins Ordinance took effect 1 August 2025 but, as detailed in Section 6, explicitly carves gold-referenced tokens out of its licensing scope — a deliberate sequencing choice that let Hong Kong stand up fiat-stablecoin regulation first while leaving room to design a gold-specific framework later as its physical vaulting infrastructure matures (AInvest, Hong Kong’s Phased Gold Integration, Jan 2026).
The most consequential regulatory non-event of the period was the EU’s continued failure to authorize a single Asset-Referenced Token under MiCA. As of January 2026, there remained zero authorized ARTs despite the category covering commodity-backed tokens like PAXG and XAUT, with national regulators still disagreeing over whether such tokens belong under MiCA’s ART regime or under MiFID II’s financial-instrument framework (MEXC, Europe Still Has Just 17 Authorized Issuers in 2026). This stands in sharp contrast to the EMT (e-money token) side of MiCA, where by March 2026 some 19 authorized EMT issuers across 11 countries were already issuing 29 e-money tokens (ESMA — Markets in Crypto-Assets Regulation (MiCA)). The divergence illustrates that gold-token issuers face a materially higher regulatory bar in the EU than pure fiat-stablecoin issuers, a gap that shows no clear sign of closing through mid-2026.
State Street launched a dedicated Digital Asset Platform in January 2026, explicitly designed to support tokenized money-market funds, ETFs, tokenized deposits, and stablecoins across both private and public permissioned blockchain networks, positioning the custody giant as an infrastructure bridge between traditional and tokenized finance rather than a gold-specific issuer (State Street, Digital Asset Platform launch, 15 Jan 2026). On the distribution side, Singapore’s DBS Bank announced plans to offer tokenized gold directly to retail customers in mid-2026, marking one of the first cases of a major traditional retail bank committing to distribute tokenized gold rather than leaving the product exclusively to crypto-native exchanges and issuers (Fidelity News (CoinDesk), DBS tokenized gold for retail, 11 Jun 2026).
Gold tokens have become increasingly embedded in decentralized finance rather than remaining purely buy-and-hold instruments. Aave governance approved adding XAUT as a supported collateral asset on its v3 Core Instance in mid-2025 and subsequently adjusted its supply cap and debt ceiling as adoption grew (Aave Governance Forum, ARFC: Add XAUt to Aave v3 Core Instance, 19 Jun 2025; Aave Governance Forum, XAUt Supply Cap and Debt Ceiling Adjustment, Dec 2025). Holders now routinely supply PAXG or XAUT to Aave as collateral to borrow stablecoins without selling their gold position, trade XAUT/PAXG liquidity pairs on Curve, and take leveraged long or short positions on gold price movements through perpetuals venues such as GMX (up to 25x leverage on an XAUT/USDC pair) and HyperLiquid (up to 10x leverage on PAXG), turning what began as a simple custodial-claim token into functional collateral across the DeFi lending and derivatives stack.
Fidelity Digital Assets’ 2026 research highlighted gold’s structural rise within global central-bank reserves — noting that gold had overtaken US dollar assets and Treasuries as a leading component of global reserves even after retracing roughly 20% from a January 2026 all-time high near $5,600 per ounce — and framed this alongside continued central-bank gold accumulation as evidence of a broader shift away from dollar-based settlement systems, a thesis the firm explicitly links to the parallel growth of tokenization and stablecoins (Fidelity Digital Assets, 6 Key Trends Shaping Digital Assets in 2026, 28 May 2026; The Cryptonomist, Fidelity: Bitcoin and Gold Are Moving Away From the Dollar, 29 May 2026).
Industry analysis published in early 2026 framed commodity and gold tokenization as a mint-on-demand, physically-verified model — an institutional user remits fiat, the issuer purchases LBMA-approved bullion and stores it in an allocated, third-party-verified vault, tokens are minted only once the physical gold is secured, and redemption reverses the flow with tokens burned against physical or cash settlement (Tiger Research, 2026 Commodity Tokenization Market Analysis, 27 Feb 2026). Broader tokenization market forecasts, including analysis projecting the total real-world-asset tokenization market toward $2 trillion by 2028 across bonds, gilts, and commodities via platforms such as JPMorgan’s Kinexys and Singapore’s Canton Network, suggest gold tokenization will likely continue growing as one slice of a much larger institutional tokenization wave rather than developing in isolation (AInvest, Tokenization Market to Hit $2 Trillion by 2028, 13 Aug 2025). The single largest swing factor for gold tokenization specifically remains regulatory: a first MiCA ART authorization, or a Hong Kong framework extension explicitly covering gold-referenced tokens, would likely unlock a new tier of regulated institutional capital that has so far stayed on the sidelines despite the sector’s market-capitalization growth and deepening DeFi integration.
Tokenized gold’s growth has run alongside, though not merged with, a separate wave of sovereign central-bank digital-currency (CBDC) development in gold-relevant jurisdictions. Kazakhstan and the UAE signed a memorandum of understanding in early 2025 to collaborate on CBDC development and virtual-asset regulation, with Kazakhstan continuing to require that crypto transactions route through its regulated Astana International Financial Center (LARA on the Block, Kazakhstan-UAE CBDC cooperation, 31 Mar 2025). Kazakhstan’s digital tenge pilot has continued to attract international recognition, winning an award at a global CBDC conference in September 2025 (The Astana Times, Kazakhstan Digital Tenge Award, 19 Sep 2025), while the UAE has continued to develop its own Digital Dirham CBDC strategy in parallel with its VARA and ADGM virtual-asset licensing tracks (UAE Government, Central Bank Digital Currency Strategy). None of these sovereign CBDC initiatives are themselves gold-backed, but they illustrate a broader regional appetite — particularly across Central Asia and the Gulf — for state-level digital-currency infrastructure that could plausibly intersect with gold-tokenization policy as both tracks mature, especially given the UAE’s simultaneous role as a physical gold-trading hub.
A concrete signal of gold tokens crossing from niche holding into mainstream DeFi collateral came via Ethena’s 2026 governance discussion to expand its USDe stablecoin backing to include PAXG and XAUT, contingent on both tokens clearing a minimum $200 million open-interest threshold across eligible exchanges — a threshold both PAXG (~$205 million) and XAUT (~$293 million) had already exceeded by early April 2026 (Ethena Governance Forum, Expand USDe Backing with Tokenized Gold, 16 Apr 2026). Proposals of this kind mark a meaningful step beyond simple buy-and-hold gold exposure: if adopted, they would make tokenized gold a direct backing component of a major synthetic-dollar stablecoin, embedding physical-gold price risk into the reserve composition of a widely used DeFi-native payment and savings instrument, and further blurring the line between “tokenized commodity” and “stablecoin reserve asset” that regulators are still working through, as detailed in Section 6’s discussion of the ART/EMT classification standoff.