Digital twins and universal tokenization
Or how to make a fragmenting financial system functionally interoperable
It was 2005. I was on the team at the SEC responsible for overseeing the investment banks. The chief credit officer at a midtown-based firm was explaining how their risk aggregation system worked. At some point, he pulled a ragged one-pager from his notebook and unfolded it. It was a maze of boxes and lines that looked like an intricate circuit diagram.
“These are the IT systems that feed into our credit risk aggregation engine,” he said. He began tapping on different lines and boxes. “We have to do manual reconciliations at all of these points because, well, not all handoffs are clean. The credit risk report that we review with you every month comes from… this.” He drew a circle in the air around the whole diagram.
Later that afternoon I met with the chief credit risk officer of a downtown-based competitor, who explained their system. There was no spaghetti diagram, no handoffs, no reconciliations. Trade booking, valuation analyses, risk measurement and aggregation, were all done in a single unified system with straight through processing.
The results and fates of these institutions would diverge over the subsequent years, reflecting the brittleness and burdens of fragmentation and the agility enabled by a unified system.
In today’s parlance, we might describe the distinguishing factor between the two as one of “interoperability” of their internal IT (or the lack thereof).
Waiting for ledger interoperability…
Single unified system are alluring. And not just for dealer banks.
Last year’s BIS annual economic report included a chapter titled, “The next-generation monetary and financial system.” It focused on tokenization, stablecoins, and the concept of a global unified ledger, reflecting the hopes and dreams that everything might one day be made interoperable at the ledger level.
Real world developments since then have moved in the opposite direction, however.
The on-chain (DLT) world is fragmenting, not converging, with the rise of new payments-focused Layer 1 blockchains. Circle launched Arc in August 2025. Within a month, Stripe launched Tempo and Tether launched Plasma. These are competing with Ethereum, Solana, and other blockchains.
Blockchains are not natively interoperable. Even the same stablecoin on different L1s aren’t interoperable, e.g., USDT on Ethereum is not interoperable with USDT on Tron. While cross-chain solutions have been developed, they are either narrow in scope, prone to hacks, or rely on an intermediary platform.
Of course, in the off-chain world, central bank settlement systems are famously non-interoperable with each other as well. This is part of the reason why cross-border payments have been so slow and expensive traditionally. But even newer systems, like Brazil’s Pix, India’s UPI, and the Fed’s FedNow remain non-interoperable.
Several initiatives have been launched seeking to approximate interoperability between ledger systems — e.g., Global Layer One, Regulated Liability Network, and the BIS Innovation Hub’s Project Agora and Project Rialto. None has been able yet to advance much beyond proofs of concept, however.
What all of these efforts have in common is a desire to achieve, or at least approximate, interoperability of ledgers. The BIS report noted above emphasized that different architectures for a unified ledger exist along a spectrum:
One of these architectures might emerge in time, but ledger interoperability has, thus far, proven to be quite elusive. Current indications are that it is likely to remain so.
Asset interoperability
All is not lost, though. Interoperability can be achieved another way, specifically at the asset level through digital twins and a well-designed underlying transaction platform.
What is a digital twin?
Imagine making a digital copy of your ownership of 500 dollars. That copy is not just any copy, though. It has a particular structure with particular fields. Those fields are carefully constructed so they can be applied to a wide range of assets, from bank deposits to Treasury bills and bonds, euros and yen, equities and corporate bonds, bitcoin and ether, stablecoins, tokenized deposits, interest rate futures and equity options. Your ownership in any or all of them can be copied and recorded as a digital twin using the same template. The digital twins represent real time ownership of the underlying asset.
Now imagine there is a transaction engine. In addition to storing the digital twins, it connects via APIs to the relevant real world custody/transaction/settlement systems for each underlying asset. This ensures that whatever happens to the digital twin happens to the original asset, and vice versa, automatically. In other words, it effectively synchronizes the real world asset and its digital twin. (Note, synchronization is one of the focal areas of the Bank of England’s Project Meridian and Digital Securities Sandbox.)
So, when the transaction engine makes a change in ownership using the digital twin — say, when you spend 100 of those 500 dollars — it syncs with the relevant existing real world systems, so your real world deposit falls by 100 dollars. Automatically, every time. Click the button below for an interactive demo explaining how digital twins work.1
Why go to all this trouble?
The consistent, structured format of the digital twin functions like a standardized shipping container — it can hold a wide range of assets and its standardized form enables easy handling by the transaction engine. This means that in the digital twin world there aren’t any handoffs or reconciliations — it is a unified system. In this sense, all digital twins are interoperable. And since all digital twins are synchronized to their corresponding real world asset, this makes the underlying assets functionally interoperable.
With functional interoperability, one can break free of the limitations imposed by reconciliation and can focus on the capabilities of the transaction engine, namely, programmability. In a unified system, programmability can be operationalized with high reliability at low cost, since there’s nothing to reconcile.
A practical solution
Asset interoperability thus offers bankers, assets managers, digital asset platforms, and corporate treasurers a practical way to navigate today’s fragmenting financial system. Instead of waiting for convergence, picking a winner, or creating a manual workaround, the digital twin solution gives users visibility and control over all financial assets, including fiat currencies, stablecoins, and cryptocurrencies, through a single pane of glass based on a unified system. The only account reconciliations that need to be done are at the level of custody and settlement systems, where that has to occur anyway as part of finality.
Not many companies offer this solution. I first learned of digital twins when attending a gathering in California and meeting the founder of a fintech called Knova. In doing a quick Google search, I came across their whitepaper. I recall reading it and thinking, “They operationalized universal tokenization!” — a topic I had pursued when I was Acting Comptroller, going so far as to focus the agency’s annual symposium on the subject.
When I met Knova’s founder, Natalya, I learned that Bill Dudley, former New York Fed president, was an advisor, and that the board chairman was Marty Chavez, former CFO of Goldman Sachs and creator of the unified system (SecDB) referenced in my opening story. I joined Knova as an advisor several months later.
Tokenization more generally
The digital twin setup described above can be seen as a form of universal tokenization. It mirrors a financial asset, representing it as a digital object and uses APIs to ensure that transactions involving the digital object are synchronized in reality. Knova’s template — its shipping container — is intentionally broad to enable the mirroring of as many assets as possible. It doesn’t rely on using a blockchain as its ledger, enabling higher speed and throughput, more control and flexibility, and lower costs.
Most tokenization efforts are blockchain based, however, and they warrant discussion. Two commercial initiatives, in particular, are worth exploring in some detail.
The first is DTC’s real world asset tokenization service, which the SEC recently blessed through a no-action letter. The service is focused on tokenizing listed equities, which aligns with DTC’s role as the central securities depository (CSD) for equities. Operationally, DTC “converts” clients’ equities into tokens minted on a blockchain. So far DTC has not indicated which blockchain(s) it is using. (The SEC no-action letter refers to “approved blockchains.”)
The upshot is that DTC’s tokenization service creates a digital twin, like Knova’s, but DTC’s digital twins exist on a blockchain. This is good if you are fluent in blockchains and want to transact using crypto wallets. If, however, you are agnostic about blockchains and don’t have a crypto wallet, this might seem like an extra layer for little benefit. More importantly, the lack of ledger interoperability between blockchains and on- and off-chain systems means that users will need to manage the fragmentation with other systems themselves or pay an intermediary to do so.
The second tokenization initiative worth highlighting is Blackrock’s BUIDL fund, which is a tokenized money market fund. With BUIDL the token holders have an equity interest in a British Virgin Island’s limited company, which in turn holds the underlying assets. BUIDL token holders have an enforceable legal claim, not just a digital copy in their wallet. The legal claim is specific, though, and runs to the BVI limited company, not the underlying assets themselves.
This legal arrangement is unpacked and compared to others in this excellent survey of the legal structures of tokenized assets by Xavier Lavayssiere.
Under Lavayssiere’s taxonomy, the BUIDL fund is “complete, indirect” tokenization, while Knova and DTC’s digital twins are “incomplete” because they act merely as digital representations, not enforceable legal claims. (For Knova, the incompleteness is a feature, not a bug, as legally enforceable ownership lives only in the native system of the asset.)
Stepping back
The pace and scope of innovation in financial services today is both exciting and a bit nerve wracking, especially in the context of AI. One consequence of so much innovation is a multiplicity of choices and options. In time, things will settle and network effects and standardization will drive consolidation. Until then, though, bankers, treasurers, and others needing to manage a mix of traditional and digital assets will find themselves spending too much time piecing things together or spending too much money and giving up too much control for someone else to do so.
Digital twins and asset interoperability provide an alternative — a way to see and control a wide range of assets in one place, through one system. To some, the layering of a unified system on top of a web of non-interoperable ledgering and settlement systems may seem inelegant. But it is practical and, like hybrid vehicles, allows users to continue to take advantage of legacy systems (gas stations) and frontier systems (EV) from a single control pane, while they wait for the future unified ledger (fully EV) world to come.