A treasury holding three stablecoins has three token balances. It may still rely on one reserve custodian, a shared group of banking partners or the same route for turning tokens into spendable bank money. Diversifying ticker symbols does not necessarily diversify the process that must work when cash is needed.
The problem is a dependency map, rather than a substitute for issuer limits. Issuer exposure asks how much the institution holds against each issuer. Dependency concentration asks which underlying firms, accounts and infrastructures affect several holdings at once. Both views are needed to understand the liquidity available during a disruption.
Separate the jobs performed by a bank
A reserve bank can hold deposits backing a stablecoin. A custody bank can hold securities used as reserve assets. A settlement bank can move cash between an issuer and its customers. One institution may perform several jobs, but the risk attached to each job differs.
A custody relationship does not automatically mean the reserve securities are an unsecured deposit exposure to the custodian. A payment relationship does not mean all reserves are held there. Combining these roles into a single percentage produces a concentration number with no clear interpretation.
Circle's published reserve information distinguishes bank deposits, short-dated government assets and other reserve categories. Its bank-facing description identifies reserve management and custody relationships. Those disclosures are inputs to a map, not grounds for assuming that every stablecoin shares the same arrangements or that every operational dependency has the same legal exposure.
Build a layered view of each holding
The first layer is the legal issuer and the holder's actual redemption access. Some institutions redeem directly; others depend on an exchange, broker or authorised intermediary. The next layer identifies how the issuer holds reserves and who safeguards them. A further layer records how cash can move out and reach the institution's intended account.
Network dependencies sit alongside those layers. Two tokens issued by different companies can share a chain, bridge, custodian or wallet service. Those are not reserve-bank exposures, but they can interrupt access to both balances. A useful map preserves the distinction while allowing a common-failure scenario to traverse them.
Our article on stablecoin distribution and redemption access explains why a nominal redemption promise and the treasury's own route are different. The concentration analysis should start with the route actually available to the account being assessed.
A hypothetical portfolio that looks diversified
Assume a hypothetical treasury holds $12 million split equally between stablecoins A, B and C. Its issuer limit allows $4 million per issuer, so all three pass. Disclosures and provider confirmations show that A and B use bank X for an important cash-out step, while C reaches the treasury's bank through another route.
This does not establish that $8 million is a deposit at bank X. It establishes that the conversion of $8 million depends on a service involving X under the assumed arrangements. A failure of that service could affect two holdings together, even if their reserves remain intact and their issuers remain solvent.
Now suppose the treasury's intermediary for C also uses X to credit the treasury's destination account. The apparent alternative has disappeared for this use case. The token portfolio is still diversified by issuer, but the tested cash-access path is concentrated across the whole $12 million.
This example is deliberately hypothetical. It shows why the exposure measure must state its unit: deposit credit exposure, custody dependency or amount requiring a shared payment route. Reporting all three as bank concentration without a definition invites an incorrect decision.
Missing disclosure is a measured uncertainty
Public reports may identify reserve categories without giving a complete list of banks or the allocation at each bank. Holdings reports can be dated snapshots. A relationship can also change between reporting dates. The map should show its observation date and distinguish disclosed facts, direct confirmations and unresolved items.
An institution should not fill a missing allocation with an assumption of equal distribution. Nor should it interpret an undisclosed bank as zero exposure. A practical report can show known minimum dependencies and an unallocated portion, with the decision governed by an explicit policy for missing information.
Treasury can request additional information from the issuer or intermediary where its relationship permits. The answer should specify whether it concerns reserves, custody or settlement services, and its effective date. A list of partner logos without role definitions adds little to the analysis.
Stress the shared service, not only the token price
A stablecoin may continue trading near its intended value while a particular institutional redemption path becomes unavailable. Conversely, a secondary market can offer liquidity when direct redemption is delayed, but only at the executable price and size available. Reserve concentration and market liquidity need separate scenarios.
One useful exercise makes a reserve bank inaccessible for a defined period without assuming a loss on all assets connected to it. Another removes a custody service while preserving ownership of segregated securities. A third removes the treasury's preferred intermediary. Each scenario asks a different question about access, timing and replacement cost.
Our article on stablecoin payments for corporate treasury provides the operating context. The concentration test should end with the payment that must be made, its denomination and deadline. A portfolio-level percentage cannot show whether a specific supplier or margin payment can still be funded.
Alternatives need independent plumbing
An alternative issuer helps only to the extent that the relevant dependency differs. An alternative redemption route helps only if the institution can actually use it under its account terms, limits and operating hours. An untested account opening or unsigned service agreement is not available contingency capacity.
Switching tokens can also use the same affected infrastructure. A conversion venue might have liquidity but depend on the treasury's existing custodian for release, or on the same bank for the final cash credit. The contingency should map the full sequence, including network access and authorisation, before counting it as independent.
The goal is not to eliminate every shared provider. Some dependencies are concentrated because infrastructure has scale or because a service has few substitutes. The decision is whether the institution understands the common dependency, accepts its consequence and has appropriate capacity for the payments that cannot wait.
Make concentration reporting answer a decision
A useful treasury report can explain which balances share a legal issuer, which reserve assets share custody and which expected cash flows share a payment route. It should show dated evidence and gaps without pretending those categories are additive dollar losses.
The most revealing question is simple: if this provider became unavailable tonight, which planned payments would fail, and what already-operational route could fund them? Answering it requires more than splitting holdings among several stablecoin names. It requires tracing the dependencies beneath each balance all the way to usable cash.