Interoperability is the missing piece, not the feature

A large US bank can now mint a tokenized deposit, settle it on a permissioned ledger, and move it between two of its own accounts in seconds. That part works. The demos are real. What that same bank cannot reliably do is send that value to a counterparty at another institution running a different ledger, under a different governance model, with a different settlement asset, and have both sides treat the transfer as final. The plumbing stops at the property line.

This is the quiet result of the last two years of bank digital-asset work. The minting problem is largely solved. The connectivity problem is not, and it is the one that determines whether any of this matters.

Islands, not a network

Look across the larger US institutions and a pattern emerges. JPMorgan has Kinexys and its deposit token. Goldman is positioning GS DAP to be industry-owned rather than captive. Several of the largest banks are coordinating tokenized-deposit work through The Clearing House. Regional efforts are standing up their own permissioned environments. Each of these is competent. Each is also, structurally, an island.

The strategic logic is sound. Tokenized deposits, not stablecoins, are the big banks' preferred instrument precisely because they keep the liability inside the regulated perimeter. A deposit token is a defensive move. It lets the institution offer programmable, on-chain settlement without ceding the customer relationship or the deposit base to a stablecoin issuer. From a risk and supervisory standpoint, that is the right instinct. Keeping the obligation inside the framework you already operate is almost always the conservative choice.

But the instinct that makes tokenized deposits attractive, the desire for control, containment, and staying the principal, is the same instinct that produces fragmentation. If every institution builds a contained environment optimized for its own control, the system of institutions ends up with a dozen contained environments that do not talk to each other. The thing that made each island safe is the thing that keeps the islands apart.

There is a telling consistency in how banks describe their own roadmaps. Ask a digital-asset team what the next hard problem is and you rarely hear "we need to tokenize more asset classes." You hear "interoperability." Connectivity. The ability to settle across institutions without forcing everyone onto one operator's rails. The market has already named the missing piece. It has been less honest about why the piece is missing.

Closed networks recreate the problem they solve

A shared network sounds like the obvious answer, and in some respects it is. The convergence of JPMorgan and Goldman work onto Canton and Digital Asset shows that institutions can agree on common infrastructure when the governance is acceptable. The trouble starts when "shared" quietly means "shared, on terms set by whoever operates the network."

A closed network solves fragmentation only for the members who join it on the operator's terms. For everyone else it is one more island, and frequently the most demanding one, because joining requires adopting that operator's ledger, its membership rules, its settlement asset, and its view of finality. The institution that joins has not eliminated a counterparty dependency. It has created a new one and concentrated it.

This is where the second-order risk lives, and it is the kind examiners now scrutinize as ordinary third-party risk management. The deregulatory shift of 2025 did not lower the bar. It moved the question into plain safety-and-soundness and TPRM terms. A bank that routes settlement through a network it does not control has to be able to answer several uncomfortable questions to a supervisor. What happens to in-flight value if the operator fails or is acquired? Is settlement finality recognized under insolvency and finality law, or only operationally fast? Are the bank's assets segregated and bankruptcy-remote, or commingled with the operator's? Who holds the keys? The FTX lesson, never commingle and never blur custody and execution, applies with full force to infrastructure, not just to exchanges. Concentration on a single network is a concentration risk, and it has to be defended with evidence, not explanations.

So the bank faces a genuine dilemma. Stay on your own island and you are fast but isolated. Join someone else's network and you trade isolation for a dependency you have to defend to your examiner every quarter. Neither resolves the problem the bank actually has.

Connectivity as a neutral layer

The way out is to stop treating the network as the product and start treating connectivity as the product. The useful unit of infrastructure is not another ledger that everyone must adopt. It is a neutral layer that lets each institution keep its own ledger, its own keys, its own customers, and its own controls, and still reach the others.

Concretely, that means a few things the banker vocabulary already covers. Settlement should be something the institution chooses: where it settles, and in which settlement asset, on its terms rather than the network's. Identity and screening should travel with the transaction. CIP and KYC at onboarding, OFAC SDN screening of both counterparties and wallet addresses including indirect, transaction-graph exposure, and Travel Rule originator and beneficiary information transmitted at the relevant thresholds, without each island reimplementing the entire BSA program from scratch. Recordkeeping should produce immutable, exportable, examiner-ready audit trails on both sides of a cross-institution transfer, so that connectivity does not become a gap in the FFIEC-shaped controls each bank already maintains. And it should integrate with what the bank runs: ISO 20022 messaging, clean reconciliation back to core, and the existing custody accounts, rather than standing up a parallel stack the institution has to govern separately.

The test for any connectivity layer is simple and slightly adversarial. Does it make the alternatives the bank's choice, or the operator's? If the layer dictates the settlement asset, the chain, or the membership terms, it is a network wearing the word "neutral." If it lets the institution decide and merely transmits between those decisions, it is connectivity. The difference is not branding. It is whether the bank remains the principal or quietly becomes a tenant.

This is the worldview behind Suave, and it is worth stating plainly so it can be argued with. The on-chain economy will be built by many institutions on infrastructure they each control, and the value will accrue at the joints, the neutral connections between regulated environments, not in any one operator's walled garden. We are early, and we are building for that layer rather than competing to be the island everyone has to move onto. That is the one claim we will make here. The rest of the argument should stand without it.

What to ask before building another island

For a strategy, risk, or compliance leader weighing the next phase of digital-asset work, the productive questions are not about which network to join. They are about what happens at the edges.

Where does our value go when it leaves our institution, and who controls that path? If a counterparty is on a different ledger, can we settle without both of us adopting a third party's chain? When an examiner asks us to defend our settlement venue, do we have segregation, key control, and recognized finality to point to, or do we have a vendor's assurances? And if the network we join consolidates, gets acquired, or changes its terms, what is our exit?

The institutions that get the next decade right will not be the ones with the most polished island. They will be the ones that kept their own ledger, their own keys, and their own perimeter, and could still reach everyone else. The minting problem was the visible work. Connectivity is the work that decides whether the visible work compounds or strands. It was never a feature to add later. It is the piece the whole thing was missing from the start.