Interoperability is the biggest challenge and opportunity in financial markets

In my view, the biggest risk for tokenization is market structure and incentives. Tech and regulation matter too, but they're easier problems to solve.

The case for tokenization usually follows this logic:

1. In most financial markets, transactions are slow to settle, often taking days, and frequently settlement fails entirely.
2. These settlement delays and risks make financial assets harder to price and financial businesses more complex and expensive to operate.
3. Today, these transaction costs are painful and costly for humans, but they’re going to be much more costly for agents, which will be significantly more numerous than humans and able to take actions and make decisions much faster than humans.
4. Blockchains can reduce transaction costs significantly for all types of transactions and make markets more efficient.

This claim makes a lot of sense when considering where transaction costs come from. There are usually several distinct ledgers for different assets involved in a transaction: commercial bank money lives inside banking ledgers; central bank money lives at the Federal Reserve, and securities live across broker-dealers, registered agents, and central securities depositories. In most financial markets, for a transaction to settle or complete, all of these participants need to update their independent ledgers, i.e., their canonical views of participants' accounts, liabilities, and assets. Whenever a transaction settles before the ledgers have been updated, a counterparty is taking on the risk that one of the participants might not be “good for it” (“it” meaning whatever asset needs to be moved out of their account).

This process should be easy to speed up because all ledgers live inside computers now, but it is not. Many ledgers cannot be updated safely on nights and weekends, often because a person needs to be in the loop. Plus, transactions are frequently batched to reduce gross liquidity requirements and/or mutualize risk. Furthermore, there are commonly more than two participants in a transaction: the counterparties, their custodians, transfer agents, banks, etc.

Blockchains offer a convenient solution for all these problems that reduces transaction costs: All of the relevant participants in a transaction can share a single “canonical” or “universal” ledger without having to trust a central ledger operator. A digital ledger with smart contracts can represent all of the relevant assets in the transaction: the commercial bank money, central bank money, and securities, and smart contracts can enable participants to settle the transaction atomically, i.e., in an all-or-nothing fashion, no matter how complicated or bespoke the underlying workflows are. Together, these three characteristics (a shared view, a single record for all assets, and programmability) make blockchains extremely valuable for settlement.

The Platform Landgrab

The only problem is that this is not the way the market is evolving. We do not have a single universal ledger or anything approaching that. Instead, it’s an all-out landgrab. The technology shift to blockchains has created credible opportunities for many different market participants to try to offer the canonical platform or network for whatever financial markets they care about. This shift is accelerated by agents, which are pulling markets towards lower latency and more programmable infrastructure across asset classes and geographies.

A new class of “web3 native” challengers have spawned that want to displace the existing networks and platforms: Ethereum, Solana, Ripple, Stellar, and the protocols building on them. These players are on their way to becoming legitimate threats to the businesses whose value is predicated on offering financial market infrastructure and networks, like credit card networks, securities depositories, and messaging networks, as stablecoin and RWA issuance has inflected. And the web3 challengers aren’t the only ones who see the opportunity: companies that own important end-customer relationships with businesses and consumers, like payment processors, banks, and broker-dealers, also see an opportunity to take back margins from the networks and platforms that aggregate and link them today. Finally, the dominant financial infrastructure providers today – the payment networks, the exchanges, and depositories – are all fighting back and launching their own blockchain initiatives.

Just to name a few:

1. Almost all of the major payment networks have announced their own ledgers: Mastercard, Swift, TCH, Stripe (vis-a-vis Tempo)
2. All of the biggest banks in the world have their own internal ledgers to tokenize their own liabilities (JP Morgan, Citi, HSBC, Wells), and they’re working on offering services to clients and other banks on top of them.
3. There are consortia of smaller financial institutions anchored in different jurisdictions building their own ledgers: Canton (US-centric), Cari, Fnality, etc
4. Many central banks are developing their own ledger for central bank money, and the “central bank for central banks” (BIS) is developing a universal ledger to facilitate transactions across them
5. The biggest stablecoin issuers have their own ledgers: Circle’s Arc and Tether-backed Plasma and Stable

Blockchains have been around for almost 20 years, but the markets are still very nascent, and these players are all vying to become the universal ledger for at least one market. The end result is that at least for now, we have many “universal ledgers”, like XKCD comic:

Localized blockchain ledgers make sense for everyone involved locally; however, the promise of tokenization reducing transaction costs depends on having a universal ledger. The potential efficiency gains are a result of all participants in a market recognizing and using the same ledger for all of the assets in a transaction. This is obviously not going to happen in the way tokenization advocates imagine it.

What do we do about this?

One option: Don’t worry about it

One view in the industry is that ledger fragmentation isn’t a real problem or a threat to tokenization because markets will naturally consolidate to a small number of platforms. The view is that liquidity is the lifeblood of all markets, and players will just migrate their business to the platforms with the most liquidity over time, and the winner(s) will compound their advantages. This is especially true if having a shared ledger really does reduce transaction costs within the market. Players transacting across ledgers will be at a disadvantage.

Distinct and relatively isolated markets might end up consolidating on different platforms, and that’s totally fine if there’s little to no need for players to transact across these markets. For example, the market for US wholesale bank payments does not need to exist on the same ledger as the Japanese sports betting market.

I believe this is unlikely to play out. First, the vast majority of the dozens of banks and consortia we have spoken to emphasize that their transactions will need to involve assets issued across other ledgers, especially securities tokens issued on public blockchains and tokenized deposits issued by other banks. They’re slowly beginning to update their commercial and product strategies to be more oriented around accepting and solving this challenge as they come out of pilot phase.

Second, there are some assets that are used across many different markets. Central bank money and commercial bank deposits are relevant to almost every transaction on the planet. JP Morgan will never issue its tokenized deposits on a single network, nor will its customers be transacting on only one network. If the JPM deposit token needs to be available across many ledgers, then the transaction costs that tokenization is designed to reduce aren’t decreasing. Actually, they’re just being internalized by JPM, coordinating across all those platforms.

Finally, there was nothing technological preventing different financial institutions from agreeing to a smaller set of canonical ledgers before blockchains. There are strong commercial incentives (and often regulatory imperatives) for financial services businesses to maintain sovereign infrastructure at the account and platform layer.

Another option: Solve interoperability bottleneck

That leaves us with the challenging but more realistic option of recovering the settlement properties of a single ledger across multiple ledgers. How would you do this? The solution needs to have properties that we can derive from the ways in which blockchains can reduce settlement costs.

Properties of an interoperability solution

1. Atomic settlement: It needs to facilitate atomic settlement of assets across ledgers. If the ledgers are updated independently or if the update process might fail, we’ve reintroduced reconciliation and settlement risk. The solution needs to support all-or-nothing completion of transactions where one asset is canonically managed by one ledger and the other by another.
2. Security: It must be maximally secure. Ideally, it introduces no new threat vectors. If there is higher risk for transacting across ledgers, the participants need to price that risk and transaction costs go up.
3. Low settlement latency: It can’t introduce new settlement delays over-and-above what the underlying ledgers themselves support. If transactions don’t settle in seconds, players do not benefit.
4. Counterparty reduction: It needs to avoid introducing new counterparties, whether that’s for security, custody, or another reason. The exception is when the counterparty reduces transaction costs. For example, a counterparty that facilitates routing across multiple independent venues to optimize clearing price might actually be desirable; however, this needn’t be necessary.
5. Programmability: It needs to support the same level of workflow programmability as the underlying ledger, especially for agents. For maximum efficiency gains, we need to be able to implement capabilities like programmable escrow and provable compliance workflows into cross-ledger transactions.
6. Low fees: It needs to have no fees or fees that can be negotiated by the counterparties. If the interoperability solution introduces significantly higher gas costs or some other form of fee, it creates a completely new vector to interfere with transactions.
7. Neutrality: It needs to be open source and openly governed by its users. This is perhaps the most subtle but most important property. As soon as the interoperability protocol is controlled by one company or consortium, it can threaten the platforms and networks that are seeking to use it. Both counterparties will wonder whether the consortium or the company behind the interoperability solution will threaten their margins or their position in the value chain by favoring their interests. This is why I was so surprised to see LayerZero launch their own ledger. It directly conflicts with their interoperability value proposition.

Notice which of these properties is the hard one. Atomic settlement, low latency, programmability, and the like are engineering problems. And engineering problems can be solved. Neutrality is not an engineering problem. It's an incentive problem, and it's the same one that created the landgrab in the first place. Every credible interoperability provider eventually faces the temptation to become a platform itself, and the moment it does, its customers become its competitors.

These reasons are why the biggest risk to tokenization is not technology or regulation. Until we have an interoperability solution that no single company can direct, we won’t have true tokenization and more efficient financial markets. We’ll just have recreated the same problems with a new competitive landscape.

Delivering a neutral and credible interoperability solution is a tall order, but it’s a huge opportunity, too. Tens of billions of dollars are lost every year to high transaction costs. With AI, that’s going to scale to hundreds of billions or trillions. Reducing those costs and driving market efficiency will benefit not only banks and payment networks, but also consumers and small businesses as savings are passed down the value chain.

I’d love to compare notes if you are working on this problem or if your institution is trying to figure out how to transact across ledgers it doesn't control.

Contact us to find out more.

About the Author

Barry Plunkett is Co-CEO and Co-Founder of Cosmos, where he leads the company's work helping the world's leading institutions adopt blockchain technology to reduce costs and operate more efficiently. Under his leadership, the Cosmos stack has become the leading open source blockchain solution, giving protocols and institutions uncompromising interoperability, customizability, and security, and today it secures tens of billions in assets and settles hundreds of billions in transaction volume every year.