A physical gold bar purchase rarely resolves into a single decision. Choosing between a 1 kg bar and a 400-ounce Good Delivery bar is already a decision about which custody can hold the metal and how the bar can move later. Resolving any of these in isolation surfaces downstream — usually as documentation that fails to reconcile against the holding, or as a resale path the format has foreclosed. Three connected decisions sit underneath the purchase: format, channel, and custody. A counterparty risk filter sits above them and determines whether they reach settlement at all.
Decision One — Bar Format
Format is the first decision because every later decision sits on top of it. The format choice — bar weight and refinery origin — determines how the metal can be held under preserved audit evidence and how granular partial exit can be. It also fixes the premium structure and shapes the resale path. Two formats carry institutional weight: the 1 kg cast bar and the 400-ounce London Good Delivery bar. Smaller cast bars and intermediate formats — 100 oz, 500 g, fractional kilobars — exist for narrower use cases; the wider range falls inside the gold bar formats collection.
The 1 kg cast bar
At approximately 32.15 troy ounces and refined to 999.9 fineness, the 1 kg cast bar is the kilobar format the institutional layer standardizes on for divisible allocation. The bar is cast rather than minted, and Heraeus and Argor-Heraeus are among the LBMA-listed refiners whose 1 kg gold bars meet the format. Each bar carries an individual serial number and refiner hallmark, with an accompanying assay certificate.
Granularity is what the buyer is actually choosing. A kilobar position can be liquidated bar by bar without disturbing the rest of the holding, while the same allocation in 400-ounce bars commits the buyer to all-or-nothing exit on each unit — partial liquidation requires either a unit-level sale or downsizing that takes the bar out of institutional clearing format. Where the use case involves tranching exposure for a family office, or releasing portions against specific receivables for a treasury function, the kilobar matches the shape of the holding to the shape of the planned exits.
What the buyer pays at acquisition is partly the option to exit in pieces. The 1 kg’s per-gram premium runs higher than the 400-ounce format and lower than retail and minted formats; the gap is narrowest for buyers planning a single liquidation event, since they are paying for an exit option they will not use. The mechanics from price quote through allocation record are documented in the institutional process for buying 1 kg gold bars.
The 400-ounce London Good Delivery bar
In the 350–430 troy-ounce range and refined to a minimum 995.0 fineness, the 400-ounce London Good Delivery bar is the format the wholesale gold market clears in. The London Bullion Market Association maintains the Good Delivery List of refiners whose bars are accepted across the recognized vault network, and Heraeus and Argor-Heraeus sit on that list. A bar produced inside the list, stored inside the network, and transferred under chain-of-integrity rules trades at the institutional reference premium without reassay.
The chain is the point. Once a bar leaves the recognized vault network — moved to a vault the LBMA does not recognize, or transported under custody arrangements that break the audit linkage — return to Good Delivery status takes reweighing and reassay. The reassay cost itself is small in absolute terms. The friction it adds to liquidity matters more, since re-listing introduces calendar time the institutional clearing format normally compresses to days. A 400-ounce bar choice is, in operational terms, a commitment to stay inside that chain — which constrains custody to the Good Delivery vault network and forecloses the kind of physical relocation that smaller formats allow.
For the buyer profile that fits the 400-ounce format, scale is the gating constraint. A bar is approximately 12.5 kilograms, and at that unit size the format serves central-bank reserves and large corporate treasury allocations — settings where unit size matches position size and where per-gram premium savings outweigh the operational cost of moving in 12.5-kilogram increments.
Decision Two — Transaction Channel
Who the buyer contracts with for the metal is the second decision. The answer changes the substance of what the buyer ends up holding. Two structural models cover most institutional purchases — the bullion-bank channel and the direct physical counterparty. The choice determines whether the position is a credit claim on a bank’s metal pool or a title to named bars in a third-party vault, and what the documentary trail will look like under audit.
Bullion-bank channel (account-based)
The bullion-bank channel runs through banks operating metal-account services for institutional clients and clearing through the LBMA system. Default account type is unallocated: the buyer holds a credit balance denominated in fine ounces, while the bank holds the metal — on its own balance sheet — inside a fungible pool used to clear obligations across the market. Settlement is fast and integrates with the buyer’s existing banking relationship; what the buyer receives at the end is an account statement rather than a bar list.
Title is the trade-off. Unallocated metal makes the buyer an unsecured creditor of the bank for the value of the position. The metal in the pool belongs to the bank until the buyer takes physical delivery — with an allocation fee and bar selection — or converts the balance to allocated metal inside the bank’s custody framework, which removes the credit exposure while keeping the metal inside the bank’s operational perimeter. A treasury function comfortable holding bank-credit exposure on its gold position, and using the same bank for FX, financing, and broader treasury services, fits the channel. A function whose mandate requires the gold itself to sit outside bank counterparty risk does not.
Direct physical counterparty (refinery-origin)
A direct physical contract bypasses the account layer. The buyer contracts with a counterparty that arranges supply from a named refinery — Heraeus, Argor-Heraeus, or another LBMA-listed refiner — and the transaction settles against specific bars. Each bar carries a serial number, refiner hallmark, and assay certification; at title transfer the bars are allocated to the buyer by serial number, either in a third-party vault contracted under the buyer’s name or by physical delivery.
What the buyer ends up holding is the bars themselves. Title sits with the buyer from settlement onward, the metal appears on the buyer’s balance sheet, and the counterparty’s solvency stops mattering against the position. The documentary trail is bar-level — contract, bar list, refinery certification — and the destination decides the last document: a vault allocation record for stored holdings, or delivery documentation for delivered bars. An auditor verifies the holding against named bars in a named vault.
On the operational side, the model is slower than account credit at a bank and depends heavily on the counterparty’s refinery access and vault contracts; documentation per transaction runs heavier. The structural comparison between bullion-bank and direct-counterparty channels takes the trade-off into detail.
Decision Three — Custody and Movement
Once settlement is complete, the metal has to sit somewhere. The choice determines what the holding’s audit trail looks like and what later movement requires. Two arrangements cover almost every institutional position: allocated storage in a third-party vault, and physical delivery to a location the buyer controls. Many buyers settle into vault storage and keep delivery as an open option for later.
Allocated storage in a third-party vault
Under an allocated arrangement, the buyer’s bars sit in a third-party vault, identified by serial number, under a custody contract that names the buyer as the title holder. Title sits with the buyer; the bars are not on the custodian’s balance sheet and are not pooled with other holdings. The custodian holds the metal and maintains the allocation record; for transactions Golden Ark Reserve coordinates, custody is contracted through Brink’s as the third-party operator. The audit trail runs from refinery certification through the allocation record at vault placement into subsequent confirmations against the same serial numbers — which is what makes the position verifiable externally, whether by the buyer’s auditor, by a bank doing collateral due diligence, or by a successor entity inheriting the holding.
Jurisdiction sits inside the arrangement as the next decision after format and channel. Recognized custody jurisdictions — Switzerland, Singapore, Hong Kong, the UAE — differ in their regulatory treatment of investment gold and in their proximity to destinations where physical delivery may later become operational; vault selection and the form of the custody contract are covered on the allocated-storage page. A position kept inside the Good Delivery vault network also retains chain-of-integrity status — reallocation between recognized vaults costs less in operational friction than moving metal that has already left the network.
Physical delivery to the buyer
Delivery moves the metal out of the vault, to a destination the buyer controls — a private vault, another jurisdiction, or a depositary arrangement under the buyer’s own contracts. The trigger is usually a specific case: relocating holdings closer to an operational base, or exiting a jurisdiction the buyer no longer wants exposure to.
Once the bars leave the vault, the operational profile changes. Incoterms allocate cost, risk, and customs responsibility between shipping and receiving sides. Carrier authorization names the firm permitted to transport the bars, constrained by insurance limits and the receiving vault’s accepted-carrier list. Customs clearance follows the bullion treatment of the transit and destination jurisdictions, and that treatment is rarely uniform across them. International coverage through Brink’s reaches a broad set of destination jurisdictions, though each border still adds layers a destination figure does not capture — customs treatment, transit insurance terms, and the receiving vault’s acceptance procedure for incoming bullion.
What survives the move is the evidence chain, and gaps in that chain are the most common failure mode. Bar serial numbers and refinery certification must reach the destination intact and must be re-recorded at the receiving custody point under the buyer’s name. Bars arriving without matched documentation, or with documentation that cannot be reconciled against the original allocation record, collapse the audit trail the original purchase was structured to produce. The physical gold delivery page treats release instruction and carrier authorization mechanics in detail.
Counterparty Risk and the Evidence Trail
Format, channel, and custody are decisions the buyer makes about what the transaction looks like. Whether it takes place at all sits behind a different question. The counterparty pair has to pass the screening that the supply chain and applicable regulation require. And the transaction has to produce the documentary record that audit and any future resale will rely on. This is the filter on top of which the three earlier decisions resolve. If it does not pass, none of them are reached.
AML/KYC and source-of-funds review
If a counterparty does not clear AML/KYC, none of the other decisions reach settlement. The requirement applies to every transaction regardless of size or familiarity. An institutional buyer’s typical file covers corporate registration documents, beneficial-ownership disclosure, identification of authorized signatories, source-of-funds evidence proportional to the transaction size, and where applicable, tax-residency confirmation. The file is rarely complete on first request, and missing items extend the calendar.
What institutional buyers most often underestimate is source-of-funds review. The standard is traceability: the funds in the transaction have to connect to identifiable economic activity — operating revenue, audited reserves, a documented asset sale, named investor capital. For a corporate entity working from financial statements, the mapping is usually clean. A family office or private investment vehicle typically needs additional documentation tying funds to the underlying source, and the AML and KYC requirements page describes what that looks like. Calendar slippage usually comes from this layer rather than from corporate records.
Sanctions screening and supplier-restriction layers
Sanctions screening sits on top of AML/KYC and operates across three aligned layers. First: international sanctions screening — OFAC, EU, UK, UN, and additional regimes as applicable — run through Refinitiv World-Check or an equivalent tier-one screening provider. Second: supplier-side restriction. The agreement under which Heraeus and other LBMA-listed refiners supply metal carries its own conditions on permitted end counterparties and destination jurisdictions, operating independently of public sanctions lists. Third: the LBMA Good Delivery framework itself, whose supply-chain standards bind the refiner-to-buyer chain.
Because the layers reinforce each other rather than duplicate, every transaction has to clear all three to settle inside the institutional supply chain. A counterparty cleared by public sanctions screening may still fail the refiner’s supplier conditions, and the supply chain’s own conditions can be tighter than what public screening alone catches.
What the qualified counterparty is actually buying is the alignment itself. The assurance is that the metal carries a supply chain that survives the due diligence their auditor and their bank, plus any counterparty doing diligence on the holding, will run against it. The actors inside that chain — refiner and vault operator, together with the verification layer behind both — are named publicly on the partners and service providers page, which is what allows the alignment to be verified externally rather than taken on assertion.
The evidence set the transaction produces
What the transaction leaves behind, once settled, is a set of documents that defines the buyer’s position and supports every downstream interaction the position will have. The core record runs across five components:
- the purchase contract, specifying parties, terms, and bar parameters
- the AML/KYC file as accepted at onboarding
- payment confirmation tying funds movement to the contract reference
- the allocation record naming the specific bars by serial number
- either vault placement confirmation or delivery documentation, depending on the custody decision
Each component has its own downstream consumer. The auditor reconciles the contract and allocation record against the balance-sheet entry. The buyer’s bank examines the AML/KYC file when the gold position appears in collateral or financing conversations. Payment confirmation links the acquisition to source-of-funds documentation if that is ever requested again. The allocation or delivery record gives a future resale counterparty the means to verify what they are buying without restarting the custody chain from scratch. Composition and sequencing of the document set, and how parameters match across documents, are stepped through in documentation in institutional gold transactions.
Where the decisions interlock
Format, channel, custody, and the counterparty filter are presented in sequence but they do not resolve in sequence. The decisions couple. A choice on any one of them constrains what remains available on the others.
Take the 400-ounce format: it presupposes a Good Delivery custody arrangement and forecloses the kind of distributed physical storage smaller bars allow. A bullion-bank channel resolves the custody question inside the bank’s own framework before the buyer has explicitly chosen vaulted versus delivered, so the channel and the custody decision are made together even when they look separate on a checklist. A buyer who wants the option of moving partial holdings between jurisdictions later is, in effect, deciding format and custody at acquisition — kilobar plus allocated vault storage inside the Good Delivery network — because that combination keeps the option open at acceptable cost.
Underneath all this is the compliance filter, and whether the combination resolves at all depends on whether the counterparty pair clears it. A counterparty pair that cannot clear the three screening layers does not reach a format choice. A source-of-funds file that cannot be assembled extends the timeline regardless of how cleanly the format and custody are specified. Buyers who run documentary work in parallel with format and custody discussions close transactions faster than buyers treating compliance as the last step.
In practice, the buyer rarely makes four separate decisions and then transacts. The decisions arrive together, in the conversation with the counterparty. The counterparty’s role is to surface the couplings — to show what a given format choice implies for custody, what the channel choice produces in evidence, and what the filter requires before any of it can settle. The transaction the buyer ends up with is the resolution of all four against the buyer’s specific use case, and the buy-side intake is where the resolution becomes a specific quote and contract.