Forward the Original Title‘Money Routers’

TL;DR

Hello,

Crypto’s killer application is already here in the form of stablecoins. In 2023, Visa did close to $15 trillion in transaction volume. Stablecoins did about $20.8 trillion in total transaction volume. Since 2019, $221 trillion in stablecoins have been exchanged between wallets.

Over the past few years, the equivalent of global GDPs has been moving through our blockchains. Over time, this capital has accumulated in different networks. Users switch between protocols for better financial opportunities or lower transfer costs. With the arrival of chain abstraction, users may not even know they are using a bridge.

One way to think of bridges is as routers for capital. When you visit any website on the internet, there is a complex network in the background, ensuring the bits and bytes that display them emerge accurately. Crucial to the network is the physical router at your home. It determines how data packets should be guided to help you get the data you need in the least amount of time.

Bridges play that role for on-chain capital today. They determine how money should be routed to get the user the most value or speed for their capital when a user wants to go from one chain to another.

Bridges have processed close to $22.27 billion through them since 2022. It is a far cry from the amount of money that has moved on-chain in the form of stablecoins. But it appears as though bridges make more money per user and per dollar locked than many other protocols.

Today’s story is a collaborative exploration of the business models behind bridges and the money they generate through bridge transactions.

Show me the money

Blockchain bridges have generated close to $104 million in cumulative fees since mid-2020. That number has a certain amount of seasonality to it as users flock to bridges to use new applications or in pursuit of economic opportunities. If there is no yield, meme token or financial primitives to be used, bridges take a hit as users stick to the protocols they are most accustomed to.

A rather sad (but funny) way to benchmark bridge revenue is by comparing it to meme-coin platforms like PumpFun. They did $70 million in fees, compared to the $13.8 million generated by bridges in fees.

The reason why we see fees staying flat even though volumes have gone up is because of ongoing price warfare between chains. To understand how they get to this efficiency, it helps to know how most bridges work. One mental model to understand bridges is to see them through the colour of hawala networks from a century back. \
Blockchain bridges are similar to hawala with portals where cryptographic signatures bridge physical separation.

Though much of what is known today about hawala revolves around its association with money laundering, a century ago, it was an efficient way to move capital. For example, if you wanted to transfer $1,000 from Dubai to Bengaluru in the 1940s—a time when the Indian Rupee was still used in the UAE—you had options.

You could either use a bank, which might take days and require extensive documentation, or you could visit a vendor in the Gold Souk. The vendor would take your $1,000 and instruct a merchant in India to pay the equivalent amount to someone you trust in Bengaluru. Money changes hands in both India and Dubai but does not cross the border.

But how does this work? Hawala is a trust-based system, operating because both the vendor in the Gold Souk and the merchant in India often have ongoing trade relationships. Instead of transferring capital directly, they may settle their balances later using goods (such as gold). Since these transactions depend on the mutual trust between the individuals involved, it requires a great deal of confidence in the honesty and cooperation of the merchants on both sides.

How does this relate to bridges? A lot about bridges operate in the same model. Instead of moving capital from Bengaluru to Dubai, you may want to move capital from Ethereum to Solana in pursuit of yield. Bridges like LayerZero enable users to lend tokens on one chain and borrow on another by helping relay messages about a user.

Presume instead of locking up assets or giving gold bars, the two traders give you a code that can be used at either location to redeem capital. This code is a form of sending messages. Bridges like LayerZero use what are known as endpoints. These are smart contracts that exist on different chains. A smart contract on Solana may not be able to understand a transaction on Ethereum. This is where oracles come into the picture. LayerZero uses Google Cloud as a verifier for transactions across chains. Even at the frontiers of Web3, we rely on Web2 behemoths to help us build better economies.

Imagine the traders involved don’t trust their own ability to interpret codes. Not everybody can get Google Cloud to validate transactions after all. A different way to do this would be to lock and mint assets.

In such a model, you would lock your assets in a smart contract on Ethereum to get a wrapped asset on Solana if you were using Wormhole. This is the equivalent of your hawala vendor giving you gold bars in India for Dollar deposits in the UAE. Assets are minted in India and given to you. You can take the gold, speculate with it and return it to get your original capital back in Dubai so long as you give the gold bars back. Wrapped instances of an asset on a different chain are similar to gold bars - except that their value usually remains the same on both chains.

The chart below looks at all the variations in which we have wrapped bitcoins today. Much of these were minted in the days of DeFi summer to facilitate creating yield on Ethereum using Bitcoin.

Bridges have a few key points they can make money on:

1. The TVL - when a user comes and parks capital, that is money that could be used to generate yield. Today, most bridges don’t take idle capital and lend it out but instead capture a small portion of the transaction fees when a user moves capital from one chain to the other.
2. Relayer Fees - These are third parties (like Google Cloud in Layer Zero) that charge a small fee on individual transfers. The price is paid for verifying transactions on multiple chains.
3. Liquidity Provider Fees - This is the money that goes to individuals who park capital into the smart contracts of bridges. Presume you are running a hawala network and now have someone that moves $100 million from one chain to the other. You may not have all that capital as an individual. Liquidity providers are individuals that pool in that money to help facilitate a transaction. In return, each of the liquidity providers gets a small cut of the fee generated.
4. Mint Costs - Bridges can charge a small portion while minting assets. WBTC, for instance, charges 10 basis points for each Bitcoin

Of these, a bridge’s expense is on maintaining relayers and paying liquidity providers. It creates value for itself on the TVL from transaction fees and minted assets on either side of a transaction. Some bridges also have a staking model which is incentivised. Say you had a $100 million hawala transfer to do to a person on the other side of the ocean. You may want some form of economic guarantee that the person on the other side is good for the money.

He may be willing to gather his friends in Dubai and pool together capital to show you that he’s good for the transfer. In exchange for doing so, he may even give back a portion of the fees. This is structurally what staking is. Except, instead of dollars, the users gather around to give native tokens of the network and in exchange get more tokens.

But how much money does all of this yield? And what is a dollar or user worth on these products?

Subscribe

The Economics

The data below is slightly dirty in that not all of the fees go to the protocol. Sometimes, fees are dependent on the protocol and assets involved. If a bridge is being used primarily for long-tail assets where liquidity is low, it could also lead to the user taking on slippage for the transaction. So, while we look at unit economics, I want to clarify that the following is not reflective of which bridges are better than the rest. What we are interested in is seeing how much value is generated across the supply chain during a bridge event.

A good place to begin with is by looking at the 90-day volume and fees generated across protocols. The data looks at metrics up to August 2024, so the numbers are for the 90 days trailing it. Our assumption is that Across has higher volume due to its lower fees.

This gives a broad idea of how much money flows through bridges in any given quarter and the kind of fees they generate over the same period. We can use this data to compute the amount of fees a bridge is able to create for each dollar passing through its system.

For ease of reading, I have calculated the data as fees generated for a $10k amount being moved across these bridges.

Before we begin, I’d like to clarify that the implication is not that Hop charges ten times more than Axelar. It is that over a ten thousand dollar transfer, $29.2 of value can be created across the value chain (for LPs, relayers and the like) on a bridge like Hop. These metrics vary across the spectrum as the nature and the kind of transfers they enable are different.

The part where it gets interesting for us is when we compare it to the value captured on a protocol with that of a bridge.

For benchmarking, we look at the cost of a transfer on Ethereum. As of writing, during low gas fees, that comes to about $.0009179 on ETH and $0.0000193 on Solana. Comparing bridges to L1s is a bit like comparing your router to your computer. The cost of storing files on your computer will be exponentially lower. But the question we are trying to address here is whether bridges capture more value than L1s from the perspective of being investment targets.

Viewed through this lens and comparing with the metrics above, one way to compare the two would be to look at the dollar fee captured per transaction by individual bridges, and its contrasts with Ethereum and Solana.

The reason why several bridges capture lower fees than Ethereum is because of the gas costs incurred in doing a bridge transaction from Ethereum.

One could argue that Hop protocol captures up to 120 times more value than Solana. But that would be missing the point, as fee models on both networks are fairly different. What we are interested in is the divergence between economic value capture and valuations, as we will soon see.

5 out of 7 of the top bridges have cheaper fees than Ethereum L1. Axelar is the cheapest—at just 32% of the average fee on Ethereum over the last 90 days. Hop Protocol and Synapse are more expensive than Ethereum today. Compared to Solana, we can see that L1 settlement fees on high-throughput chains are orders of magnitude cheaper than bridging protocols today.

One way to further enhance this data would be by comparing the costs of doing a transaction on L2s in the EVM ecosystem. For context, Solana’s fees are 2% of what it would usually cost on Ethereum. For the purpose of this comparison, we will go with Arbitrum and Base. As L2s are purpose-built for extremely low fees, we will take a different metric to benchmark economic value—that of average daily fees per active user.

In the 90 days for which we took the data for this article, Arbitrum had 581k average daily users and created $82k in fees on an average day. Similarly, Base had 564k users and generated $120k in fees on an average day.

In contrast, bridges had fewer users and lower fees. The highest among these was Across, with 4.4k users generating $12k in fees. From this, we estimate that Across creates $2.4 per user on an average day. This metric can then be compared with how much Arbitrum or Base produces in fees per active user to gauge the economic value of each user.

The average user on a bridge is far more valuable than one on a L2 today. Connext’s average user creates 90 times the value a user on Arbitrum would. This is a bit of an apples-to-oranges comparison because doing bridge transactions on Ethereum comes with its share of gas costs which can be prohibitively high, but it highlights two clear factors.

1. Money routers like bridges today are probably one of the few product categories within crypto, producing meaningful economic value.
2. So long as transaction fees remain prohibitively expensive, we may not see users going to L1s like Ethereum or Bitcoin. Users could be onboarded directly to an L2 (like Base), and developers may choose to cover gas costs. Alternatively, there could be a world where users switch only between low cost networks.

A different way to compare the economic value of bridges would be by comparing it with a decentralised exchange. When you think of it, both these primitives serve similar functions. They enable the movement of tokens from one form to the other. Exchanges enable moving them between assets, while bridges move them between blockchains.

Data above is for decentralised exchanges on Ethereum alone.

I avoid comparing for fee or revenue here. Instead, what I am interested in is capital velocity. It can be defined as the number of times capital rotates between a smart contract owned by a bridge or a decentralised exchange. To calculate it, I divide transfer volumes on bridges and decentralised exchanges on any given day with their TVL.

As expected, for decentralised exchanges, the monetary velocity is far higher as users routinely swap back and forth on assets multiple times over the course of a single day.

What is intriguing, however, is that when you exclude large L2-oriented bridges (like that of Arbitrum’s or Opimism’s native ones), the monetary velocity is not too far from that of a decentralised exchange.

Perhaps, in the future, we will have bridges that keep caps on the amount of capital they take and instead focus on maximising yield through increasing capital velocity. That is, if a bridge is able to rotate capital multiple times over the course of the day and pass on fees to a limited subset of users that have parked capital, it will be able to generate higher yield than alternative sources within crypto today.

Such bridges will probably see stickier TVL than conventional ones, where scaling parked sums of money leads to lower amounts in yield.

Are Bridges Routers?

Sourced from Wall Street Journal

If you think VCs rushing to “infrastructure” is a new phenomenon, take a walk down memory lane with me. Back in the 2000s, when I was a wee little lad, much of Silicon Valley was hyped about Cisco. The logic was that if the amount of traffic going through internet pipelines were to increase, routers would catch a substantial portion of the value. Much like NVIDIA today, Cisco was a highly-priced stock as they built the physical infrastructure that enabled the internet.

The stock peaked at $80 on 24 March 2000. As of writing, it trades at $52. Unlike many dotcom stocks, Cisco never recovered. Writing this piece in the midst of a meme-coin mania made me think about the extent to which bridges can capture value. They have network effects but could probably be a winner-take-all market. One that is increasingly trending towards intents & solvers, with centralised market makers filling orders in the back end.

Ultimately, most users don’t care about the extent of decentralisation of the bridges they use. They care about cost and speed.

In such a world, bridges that emerged in the early 2020s could be similar to physical routers that are closer to being replaced by intents or solver-based networks that are closer to what 3G was for the internet.

Bridges have reached a level of maturity where we are seeing multiple approaches to the same old problem of moving assets across chains. A leading driver for change, is chain abstraction - a mechanism of moving assets across chains such that the user is blissfully unaware of ever having moved assets. Shlok recently had a taste of it with Particle Network’s universal accounts.

A different driver for volume, would be products innovating on distribution or positioning for driving volume. Last night, while exploring meme coins, I noticed how IntentX is using intents to package Binance’s perpetuals markets onto a decentralised exchange product. We are also seeing chain specific bridges evolving to be more competitive in their offerings.

Whatever be the approach - it is evident, that much like decentralised exchanges, bridges are hubs for large sums of monetary value to flow through them. As a primitive, they are here to stay and evolve. We believe niche specific bridges (like IntentX) or user specific bridges (like the ones enabled by chain abstraction) will be the primary drivers for growth within the sector.

One bit of nuance Shlok added while discussing this piece is that routers in the past never captured economic value in proportion to how much data they passed. You could download a TB or a GB, and the Cisco would make just about as much money. Bridges, in contrast, make money in proportion to the number of transactions they enable. So for all intents and purposes, they may have different fates.

For now, it is safe to say that what we see with bridges and what happened with physical infrastructure for routing data on the Internet rhymes.

Disclaimer：

1. This article is reprinted from [decentralised.co]. Forward the Original Title‘Money Routers’. All copyrights belong to the original author [Joel John]. If there are objections to this reprint, please contact the Gate Learn team, and they will handle it promptly.
2. Liability Disclaimer: The views and opinions expressed in this article are solely those of the author and do not constitute any investment advice.
3. Translations of the article into other languages are done by the Gate Learn team. Unless mentioned, copying, distributing, or plagiarizing the translated articles is prohibited.