Self-Custody Must Survive - Why Institutional Custody Threatens Blockchain’s Core Value Proposition

Earlier this year, the U.S. Securities and Exchange Commission changed how banks account for stablecoins on their balance sheets. Before the rule change, if an institution held $100 million in stablecoins, regulators treated that holding as worth nothing. Zero. After the change, that same holding counts at roughly $98 million—a 2% haircut instead of 100%.

This is not a minor accounting adjustment. It is a green light. It means every major financial institution now has a reason to hold stablecoins, to build on blockchain rails, to move money at the speed of the internet rather than the speed of correspondent banking. Combined with the passage of the GENIUS Act and the U.S. Treasury’s estimate that $6.6 trillion in bank deposits are at risk of migrating to blockchain-based accounts, the direction is unmistakable. Institutional adoption of digital assets is no longer a question of if. It is underway.

But there is a question that almost no one is asking, and it may be the most important question of this entire transition: when trillions of dollars move onto blockchain infrastructure and banks become the primary custodians of those assets, what happens to the transparency that made blockchain worth building in the first place?

The Glass Bottom

Blockchain’s fundamental innovation is not speed. It is not cost reduction. It is not programmable money, or tokenisation, or any of the other capabilities that dominate conference agendas. The fundamental innovation is transparency—radical, structural, architectural transparency.

Think of it as a glass bottom. For the first time in financial history, we built infrastructure where every participant can see the ledger. Every transaction is recorded publicly. Every movement of value is independently verifiable. Any counterparty, any regulator, any insurer can confirm the state of affairs without requesting permission, without trusting an institution’s internal books, without waiting for an audit.

This is not a feature. It is the reason blockchain infrastructure exists. Open reconciliation—the ability for any party to independently verify what has actually happened—is the property that justifies the cost, the complexity, and the disruption of moving financial infrastructure onto distributed ledgers. Without it, blockchain is just a slower, more expensive database.

How Institutional Custody Covers the Glass

When a bank adopts stablecoins under the current custodial model, its customers do not receive wallets. They do not receive private keys. They do not receive public addresses on a blockchain. They receive a bank account—the same instrument they have always had, denominated in a new type of asset.

The bank holds stablecoins in its own wallet infrastructure—typically pooled omnibus wallets that aggregate the holdings of thousands or tens of thousands of customers into a single on-chain address. On the blockchain, you see one wallet with one balance. The individual client relationships, the sub-allocations, the internal ledger entries that map specific holdings to specific customers—all of that happens off-chain, inside the institution’s proprietary systems. Invisible to anyone outside.

Even Coinbase, which is further along than most banks in this regard, assigns each user a generic deposit address that functions as a routing mechanism into their internal ledger. It is not the user’s wallet. It is a deposit funnel. The user never signs a transaction. The user never holds a key. The user’s on-chain identity does not exist.

The glass bottom has been covered with an opaque floor. The boat still floats. The ocean is still there. But no one can see through anymore.

Open reconciliation disappears. Independent verification disappears. The ability for a counterparty to confirm what is actually held, what has actually moved, who has actually transacted—gone. We are back to trusting the institution’s internal books. We are back to audits. We are back to the model that preceded blockchain entirely.

The Idiosyncratic Risk of Opacity

Blockchain was supposed to make financial crime harder to commit and easier to detect. Transparent ledgers. Traceable flows. Every transaction visible. The premise was compelling: if everyone can see the ledger, fraud has nowhere to hide.

But institutional custody recreates precisely the opacity that enabled every major financial scandal of the last three decades. The mechanics that allowed HSBC to process cartel money. The black boxes that let Wirecard fabricate billions in revenue. The commingled accounts that enabled FTX to lose customer funds without detection. These failures did not occur because of insufficient technology. They occurred because the systems were opaque, and the people operating them exploited that opacity.

If the future of blockchain-based finance is trillions of dollars flowing through institutional omnibus wallets, we have not solved this problem. We have migrated it to more efficient infrastructure. The financial crimes do not stop—they move behind the custodial wall where the glass bottom cannot reach. This is not systemic risk that can be modelled and hedged. This is idiosyncratic risk—concentrated, opaque, and specific to every institution that covers the glass.

Regulators, it should be noted, pushed for blockchain adoption in part because it offered better auditability than traditional finance. But if institutional custody negates that auditability, regulators gain settlement speed and lose the transparency dividend. That is not a good trade.

The Closed Door

There is a practical dimension to this problem that extends beyond transparency and into basic functionality. Bank accounts do not have private keys. They do not have public addresses on a blockchain. When a bank custodies stablecoins for a customer, that customer has no on-chain identity. They have no address that the outside world can send value to.

So how does anyone outside the institutional perimeter interact with someone inside it? If you are a freelancer, a small business, a DAO, or an individual in an unbanked region—and your counterparty’s assets are locked inside a bank’s custodial infrastructure—there is no door. There is no address to send to. There is no permissionless way to transact.

The promise of 24/7 open settlement collapses at the boundary. It is only 24/7 between institutions that have agreed to interoperate. For everyone outside that perimeter, blockchain-based banking is indistinguishable from traditional banking—except more expensive to operate.

Self-Custody as Structural Necessity

Self-custodial wallets are the only wallets where on-chain activity represents real, individual, verifiable activity. When a person holds their own keys and signs their own transactions, the glass bottom works as designed. Every transaction is attributable. Every movement is traceable. Every counterparty can independently verify what happened without requesting permission from a third party.

Self-custody is also the only model where an individual has an addressable identity on a public blockchain—an address that anyone can send to, interact with, and verify. Without it, there is no permissionless participation. There is no open settlement. There is no glass bottom.

This is not an ideological argument. I am not making the case for self-custody because of libertarian principles or cypherpunk philosophy. I am making the case because self-custody is the only model that preserves the one property that makes blockchain infrastructure worth the investment: verifiable, transparent, open financial activity. If self-custody dies, the glass bottom shatters, and blockchain becomes an expensive recreation of the system it was built to replace.

The Privacy Paradigm

I want to be honest about a real tension in this argument. Total transparency has its own problems. If you operate from a single wallet on a public blockchain, anyone with your address can see your balance, your entire transaction history, every counterparty you have ever interacted with. In traditional banking, if someone has your account and routing number, they can send you money and confirm whether you can cover a specific amount—a simple yes or no. They cannot see your balance. They cannot see your history. The glass bottom, pointed at an individual with no structural privacy, is not an improvement. It is a different kind of exposure.

But the answer to this tension is not to abandon transparency and rebuild the same opaque banking system on blockchain rails. That solves nothing. It preserves nothing. It justifies none of the investment.

What I believe is this: at some point, we need infrastructure that exposes enough of what we do to hold the institutions responsible for our money and our reputation to at least the same standard—or a higher one. My transaction history, available as a foundation for my identity, is fundamentally better than handing a photograph of my passport to twelve different platforms and hoping none of them get breached. And the assumption that a single wallet must equal a single financial identity—that every aspect of my financial life must be visible in one place—is itself the constraint. It does not have to work that way.

What Self-Custody Lacks

The institutional objection to self-custody is legitimate. Self-custodial wallets today have no identity layer. A wallet address is a string of hexadecimal characters. It tells you nothing about who controls it, whether they have been verified by any institution, whether they are creditworthy, or whether they are even a real person. Banks and regulators look at a self-custodial wallet and see a black box—which is, ironically, the same complaint they make about each other.

There is no proof of who connected. No verification of institutional relationships. No auditable record that an authentication event ever took place. This absence of an identity and verification layer is the reason institutions do not trust self-custodied wallets, and it is the reason the default path is institutional custody—with all the opacity and all the idiosyncratic risk that entails.

The gap is not self-custody itself. The gap is the absence of trust infrastructure that makes self-custody legible to institutions without destroying the transparency that gives it value.

What Gets Built Next

Trillions of dollars are moving onto blockchain infrastructure. The regulatory signals are clear. The institutional appetite is real. This is happening.

The default path is institutional custody—and with it, the return of opaque ledgers, concentrated risk, and a financial system that looks remarkably like the one blockchain was supposed to replace. That path is well-funded, well-understood, and already underway.

But consider what institutional custody actually protects. Banks that commit crimes with customer funds can do so because they hold those funds. Banks that overleverage deposits into insolvency can do so because they control those deposits. Banks that fail and take ordinary people’s savings with them can do so because they possess what was never theirs to risk. Every one of these failures is structurally impossible when customers hold their own keys.

Self-custody must survive. Not as a niche preference. Not as an ideology. As the structural foundation of a financial system that is transparent, verifiable, and accountable. The trust infrastructure to make it work at institutional scale does not fully exist today. But the need for it is no longer theoretical.

The question is no longer whether self-custody should survive. It is how we make it happen.