Custody wins the tokenization back office

The CFTC's move to put tokenized collateral behind an equivalence proof turns a sales pitch into a compliance obligation, because customer funds can sit in tokenized assets and a distributed ledger can be the official record, but the wrapper must now prove the token is as safe as the permitted assets it replaces. That shift changes the buyer from a treasurer who was sold a token to one who must be shown an equivalence certificate, and the firms that can supply that certificate—or distribute the custody behind it—get paid before a single token is issued. This is a back-office entry, rewarding the certifiers before the token market opens.

Regulators have historically treated tokenization as a securities or settlement question, but this CFTC posture makes it a custody question first by allowing the ledger to be the official record—the largest concession the digital-asset industry could have asked for—while making that concession conditional on the wrapper proving equivalence to the assets customer funds are already permitted to hold. That is where the economics move. The proof is a standing requirement every custodian, venue, and treasury team must meet every time a tokenized asset is offered against customer funds.

IBM's Swift beta shows what distribution without settlement looks like: banks can instruct a tokenized deposit over the payment messages they already send, which amounts to a custody platform with an enormous pre-installed client base and no answer yet on who holds the asset. In a market about to be asked to prove safety everywhere, leaving the custody question open gives a custody provider the surface on which to later stamp its own standard, because the distribution channel is the scarce thing and IBM just borrowed Swift's. A bank that already sends Swift messages for every cross-border payment does not need a new ledger to instruct a tokenized deposit—the tokenized deposit adopts the message format the bank already uses.

Bitget's move makes off-exchange collateral a baseline, because four venues on one bank's platform turn segregated off-exchange collateral from a marketing line into a cost of doing institutional business. Once four venues share a bank's platform, no venue can sell segregation as a differentiator—the bank's platform becomes the standard and the venues become tenants—which makes this a custody business rather than an exchange business, and it is being built before most institutional treasuries have even finished their tokenization policy reviews. The venues are paying for access to a balance sheet rather than competing on infrastructure.

Bullish and Equiniti are making the register the control point by convening a standards group for register-linked tokenized equities around Equiniti, the shareholder-services firm Bullish is buying. The register is where legal ownership is recorded, so tokenizing a register-linked equity moves the ledger to the register, and whoever owns the register owns the standard that determines how the token maps to the underlying share, a piece of infrastructure most token issuers do not own. The buyer in that transaction gets the record of who owns the stock rather than a token issuance platform.

The seven UK banks moving real money across a shared tokenized rail make the interbank leg matter more than the remortgage payments, because the remortgage payment is the demonstration while the rail is the franchise. Once real money moves between banks on a shared tokenized rail, the question shifts from whether tokenized deposits work to who else gets on the rail, and the banks that build the rail before it becomes regulated infrastructure get to set the access terms, which are pricing power. A tokenized deposit that moves on seven banks' rail is a different instrument from one routed through a separate ledger, because it inherits the existing interbank settlement network.

The market-size argument is really a collateral mobility argument: a $2.3 trillion tokenization forecast sounds like a product market, but its real number is $1.7 trillion in collateral mobility, which belongs to the balance sheets and venues that move the assets rather than to the products being pitched. The remaining $600 billion of tokenized instruments is the visible market, while the hidden one is the collateral mobility pool, because collateral moves between venues, banks, and counterparties to satisfy margin and funding obligations. A treasury that posts tokenized collateral through an existing bank platform reuses the same margin call and the same custody account with a different wrapper.

The winners are not token issuers, because the token is the wrapper while the asset is the bank relationship, the existing message rail, the register, and the interbank ledger. IBM's Swift beta borrows a distribution channel banks already use; Bitget's platform turns a bank's balance sheet into the venue layer; Bullish and Equiniti own the register-linked standard; the seven UK banks own the rail; and the CFTC equivalence proof makes safety certification a moat. In each case, custody is being distributed over infrastructure someone else already runs, and the money lies in the certification and the access, leaving issuance to the side, so a token issuer that wants to reach institutional collateral must buy one of those relationships or rent one.

Most tokenization conferences are still pitching the token—a fund a customer can hold onchain, a treasury that settles instantly, a bond that clears without an intermediary—but the back office is asking which register is authoritative, which message rail carries the instruction, which bank's platform houses the collateral, and who certifies that the wrapper is as safe as the asset it replaces. Those questions are answered by the firms named above, and they are answered through existing bank rails rather than new ledgers, so the conferences are selling the product while the back office is buying the rails.

The CFTC's equivalence proof still needs a compliance template, and whoever writes it will turn a cost center into a product, just as the UK rail's first seven banks have built the leg and the next entrants will decide whether it remains a bank utility or becomes an open standard. The $1.7 trillion collateral mobility pool sits on the other side of those two decisions, so the firms that write the template and set the rail access terms will own the next decade of institutional digital assets without ever asking a customer to buy a token.