April 9, 2026
Carson Cook, CEO
/
Institutions in digital asset credit have built sophisticated internal infrastructure. The boundary between those institutions remains largely ungoverned. There is a pattern in financial history that speaks directly to what tends to happen next.
By Carson Cook, PhD, CEO & Founder, Membrane Labs
Building internal operational infrastructure is a reasonable response to a fragmented market. Most institutions in digital asset credit and collateral have done it. The internal problem — tracking positions, managing approvals, running workflows — is largely solved inside individual firms. What tends to surface later is the boundary problem: the counterparty on the other side of the trade built something different. The two systems share no common state model. The operational record of what is happening between firms still travels by email and spreadsheet, because nothing was built to carry it.
There is a structural reason why that boundary problem is harder to solve than the internal one, and it is worth understanding. The history of traditional financial market infrastructure offers a useful reference point.
The Paperwork Crisis and What It Taught Us
A broker sorting through piles of physical stock certificates, circa 1968. Image: Getty Images.
In the late 1960s, the New York Stock Exchange had a problem it had not anticipated. Trading volume was growing faster than the back office could process it. Settlement was bilateral; each firm maintained its own records, issued its own certificates, and reconciled with counterparties through a system of physical paperwork. When volume spiked, the paperwork piled up. By 1968, the exchange had to close on Wednesdays just to let the back offices catch up. Firms went under. The phrase used at the time was “the paperwork crisis.”
Over just two years, about a sixth of all NYSE member firms disappeared — through merger or closure — and almost every one had suffered badly in its back office. In the operational chaos, organized crime moved in. At a 1971 Senate hearing, the U.S. Attorney General estimated that syndicates had stolen more than $400 million in securities. The absence of a shared operational record did not just create inefficiency; it also led to errors. It created conditions that could be exploited by bad actors.
The answer was the Depository Trust & Clearing Corporation, which most in this industry now know as the DTCC. A single neutral entity that held securities on behalf of all participants and handled settlement through book-entry transfers rather than physical certificates. No more bilateral reconciliation. No more certificate movement. A shared operational record, maintained by a party that had no stake in the outcome of any individual transaction.
SWIFT followed, for a similar reason, in a different domain. Banks needed a standardized messaging network for international transfers. The neutrality was not incidental to the design. It was what made adoption possible — banks would use infrastructure that no single competitor controlled, and would not use infrastructure that one did.
In both cases, the institutions best positioned to build comprehensive infrastructure were also the institutions that counterparties would have been most reluctant to depend on. The conflict was structural rather than personal. The answer, in both cases, was an entity that sat outside the competitive relationships of the market it served.
The neutrality of the DTCC and SWIFT was not a product feature. It was the precondition for adoption.
The Same Problem. Fifty Years Apart.
Figure 1. The same structural problem — bilateral fragmentation resolved by neutral shared infrastructure — recurring fifty years apart.
Digital Asset Credit Markets in 2026
Consider what a digital asset credit workflow actually looks like today. A lender books a loan against digital asset collateral. The collateral sits at one custodian, sometimes two. The loan terms live in a bilateral agreement. Pricing comes from a third-party feed. Settlement goes on-chain. The counterparty has their own books and records that represent the same transaction in their own format. None of those systems were designed to share a common state model.
When an LTV threshold is breached and someone needs to post more collateral, determining who owes what, to which custodian, in what state, requires manual reconstruction across those disconnected environments. Phone calls, Slack messages, spreadsheets with timestamps. At small volumes, this is workable. At the volumes institutional digital asset markets now process, that manual coordination layer introduces latency and ambiguity that a shared operational record would eliminate.
FROM THE FIELD
When we started mapping the operational workflows of institutional digital asset lenders, the internal systems were often surprisingly sophisticated. The breakdown was always at the same place: the point where one firm’s system had to communicate state to another firm’s system. That gap was always manual. It was always fragile. And it looked the same across every institution we spoke to.
The internal build is a reasonable first step. What emerges later is the boundary problem. The counterparty built something different. The formats do not align. The state models do not correspond. The operational record of the transaction — the one both sides would need to resolve a dispute cleanly — does not exist as a shared artifact. It exists in parallel, in each firm’s own systems, interpreted differently.
This produces friction at exactly the moments when friction is most costly: when a threshold is breached, when a counterparty needs to act quickly, when something in the workflow has gone wrong and both sides are trying to reconstruct what happened.
Why In-House Builds Cannot Solve This
Figure 2. The boundary problem in institutional digital asset credit — and the neutral infrastructure answer.
The Trust Problem in Participant-Owned Infrastructure
The structural reason in-house builds cannot solve the boundary problem is straightforward. Any orchestration layer controlled by a market participant carries a conflict that limits adoption. Goldman Sachs will not route operational state through JPMorgan’s infrastructure. A borrower will not depend on their custodian’s workflow engine when that custodian also lends against the same collateral pool. This is not a technology problem. It is a trust problem, and better engineering does not resolve it.
In the EY/Coinbase 2026 institutional survey of 352 global institutions, 59% of respondents cited integration challenges as a top barrier to digital asset participation — second only to regulatory uncertainty.
59% cite integration challenges as a top barrier
EY / Coinbase 2026 Institutional Investor Digital Assets Survey, 352 global institutions. Second only to regulatory uncertainty (67%).
Integration Is the Solvable Problem
Figure 3. Top barriers cited by institutional investors when investing in or holding digital assets. Integration challenges rank second only to regulatory uncertainty — and unlike regulation, they are solvable with infrastructure.
The DTCC and SWIFT were not primarily technology solutions. They were governance answers to a trust problem. The DTCC works because it does not trade. SWIFT works because it does not move money. The clearinghouse works because it is not a counterparty. The neutrality of the operator is what allows the operational data to flow between competitors.
Digital asset credit and collateral markets are facing a structurally identical problem. The question is whether a similar structural answer will emerge, and when.
A Note on Settlement Infrastructure Specifically
There is a dimension of this that is specific to digital assets and worth mentioning separately.
On-chain settlement infrastructure offers capabilities that traditional rails cannot straightforwardly replicate: settlement finality, programmable execution, on-chain auditability at a level of granularity that has no real precedent in conventional finance. On-chain credit and liquidity structures extend that further, with composable frameworks that do not require a central intermediary to function. Institutional interest in these capabilities is genuine and has been growing.
The adoption rate has been slower than the commercial interest would suggest. The primary friction is operational rather than regulatory. A regulated institution connecting to an on-chain settlement rail still needs audit trails, approval workflows, exposure controls, and role-appropriate visibility into positions and counterparty states. On-chain rails do not provide those. They provide settlement. What sits between a regulated institution and an on-chain rail is a compliance and control layer that each institution currently builds and maintains independently, in formats that are not compatible with each other.
If that control layer were shared neutral infrastructure, the connection between regulated institutional operations and on-chain settlement rails becomes considerably more tractable. More institutional participation would change the character of tokenized credit markets. Access to on-chain settlement efficiency would change what institutional treasury and credit operations can do. One condition precedes the other.
The Technical Question Underneath All of This
The governance and trust argument for neutral infrastructure has a technical prerequisite: the infrastructure has to be buildable.
For an orchestration layer to work across institutional boundaries, every operationally significant state transition — obligation creation, collateral posting, margin breach, substitution request, settlement instruction — needs a single agreed representation that all participating systems can read, regardless of which custodian, chain, or internal platform originated it. The term of art for this is a canonical event model: one authoritative schema that normalizes the operational diversity of institutional digital asset markets into a shared, legible, auditable form.
The difficulty is that the schema has to be expressive enough to capture real institutional controls and abstract enough to remain stable as the custody landscape evolves and new chains and venues get added. If it requires institution-specific extensions for every new counterparty or environment, it is not really canonical in the meaningful sense — it is a translation layer, which solves a narrower problem and does not generate the network effects that make shared infrastructure valuable.
Whether a genuinely canonical model is achievable at institutional fidelity is the question Membrane Labs is working to answer. We have good reasons to think it is. We are testing that rather than assuming it.
On Timing
There is an argument that markets self-organize toward interoperability eventually — that shared standards emerge naturally once the cost of fragmentation becomes visible enough. That may be right. The history of financial market infrastructure is somewhat equivocal on the point.
Protocol-level standardization tends to move slowly, often at the pace of the most reluctant participant. And the cases where shared infrastructure emerged relatively quickly — the DTCC is one example, SWIFT is another — tend to involve a triggering event that made the status quo visibly untenable. The paperwork crisis was that kind of event. It concentrated the attention of enough participants simultaneously that coordination became easier than inaction.
Digital asset credit markets have not had that event yet. The coordination problem is already present at the volumes the market currently processes, not only at some future scale. Whether the market addresses that proactively or reactively is an open question.
Membrane Labs is working on the proactive version. If the canonical event model is achievable at institutional fidelity, the infrastructure becomes possible before a crisis makes it necessary. That is the better sequencing.
THE BOTTOM LINE
The paperwork crisis was solved. The problem digital asset credit markets face now is structurally similar and technically more tractable. The question is not whether the infrastructure gets built. It is whether it gets built before the market needs it to contain a failure, or after.
Carson Cook, PhD
Carson is CEO and Founder of Membrane Labs, which builds counterparty-neutral infrastructure for institutional digital asset credit, collateral, and settlement. He holds a doctorate in nuclear physics and previously worked at McKinsey and Company. He is the named inventor on two issued U.S. patents covering coordinated digital asset settlement. Membrane Labs is headquartered in Miami, Florida.
