What is the Next Evolution for Layer 2s?

Transcript

Bonjour, hello. All right, cool. Everyone can hear me. How's it going everyone? My name is Matt. Um, hope you all are having a nice day in this incredible weather in the spring. I'm here to talk about our favorite subject, the future of programmable economies. Uh, originally and the general focus of this talk is to talk about scaling solutions for Ethereum. Uh, again, my name is Matt. I work at Offchain Labs, the initial developers of the Arbitrum protocol, which is the leading layer two in EVM. So, what are programmable economies? What are layer twos? Um, interchangeable now, but you're familiar with blockchains.

They allow you to put rules for trade in a program that a bunch of different people can come to agreement on the result of. And we're here to make those faster and cheaper and more flexible. And we're starting to see a lot of this happen in the form of customization at the chain level and customization at the account level. And there are reasons why this is accelerated the last few months and in the last year and I'm here to talk about why and what to expect going forward with all the different products that you're using now. So, what I'm going to talk about, brief overview on some interesting metrics that I think are useful to understand some of the points I want to raise.

Um, going to cover specialization and why focusing on certain transactions drive economic value. I'm going to talk a little bit about ordering at the chain level, particularly with regard to sequencing, which if you're familiar with L2s, are the software or party that orders the transactions that hit the network. And then I'm going to cover a few small points at the end that are just some interesting things to consider. So, one of the things I want to talk about is how value creation drives concentrated activity. It was just a one-line way that I could fit this into a small part of this of the slide, but you see this in the world generally.

Concentrated liquidity is a great example where there's value, people flock to. It's easier to transact with parties that are close to you. This kind of flows down into software and into blockchains. One of the things I wanted to highlight was, you know, certain financial transactions that are happening now predominantly on blockchains, not compared to the traditional markets, but compared to blockchains overall. There's really only a few transaction types that actually matter, create value, and to people use regularly. A great example is borrowing and lending. You're familiar with Aave, Morpho, other lending protocols. What you see here is just a simple chart that goes over how concentrated those deposits really are when you look at the network breakdown.

So, if you look at all of the deposits in the Aave protocol, the vast, vast, vast majority of them are all in Ethereum. Obviously, this is due to different effects, first-mover advantage, you know, iteration on the protocol, the team is great, but this is just something that compounds over time. That matters. Another example is a newer type of financial primitive we've seen over the last decade evolve, which are perpetual futures, obviously a more capital-efficient version of options contracts, very good way to have leverage on a trade, very, very popular as a blockchain application because it really is the only place where you can do this in a way that allows the system to work efficiently enough for those trades to hit with low spread. So, you see here, this is still fluctuating.

The protocols themselves are still developing, but you notice the light blue or green, um, depending on your opinion, color in the graph, that's the Hyperliquid protocol, obviously pioneers in the perp space. But we're starting to see, you know, another example of a transaction type have concentrated action on a specific chain. Now, remember, Hyperliquid is its own, you know, kind of blockchain. There's a bridge on Arbitrum, but they've established their own way of processing these transactions to fit their need. This is a great example of how this happens. Another example, you know, is Polymarket on Polygon. Right here, this is a share of gas used on the overall Polygon network from Polymarket compared to all other applications on the network. So, if you look at this, obviously this fluctuates day to day,

but in the last few months, there's this run-up where I think at a certain point at the high, Polymarket was responsible for 84 to 85% of all gas consumed on what is otherwise a general network. So, again, this is a real application with real value that people are using to do something productive, right? There's a lot less speculation going on because we have now tested the software, we've tested our assumptions, we we can kind of judge and retail can judge now what is a useful transaction versus what is not. So, >> [snorts] >> how does this matter to L2s? Well, if you have a transaction that is valuable and a lot of people are doing it, now you have limited block space, more competition. So, there's a natural trajectory for a successful application on a standalone general-purpose network to want to dedicate block space to

themselves, launch their own chain. So, what I'm going to talk about are some few steps that, you know, you might see chains doing. We've seen chains doing some of this. Alt to L1s have done this for a long time. On the range of simple to complex, pardon my hand-drawn arrows. So, I think something we're definitely going to see and we think a lot about this at the in the Arbitrum space is looking at EVM chains in particular. I think we're going to see a lot of chain owners take critical application logic and make that part of the core state transition function of the chain. You can think about this rather than having just a general-purpose smart contract layer where people put applications on, let's just make all the nodes responsible for doing payments or let's make all the nodes responsible for computing like a look-back period for perps, right? Where

you you take a look at the the historical orders and try to calculate the funding rates. Make that the responsibility of everyone in the network. And in a way, this is what Hyperliquid has already done. So, we're going to see this happen a lot more on general-purpose chains where they take the prevailing value-creating transactions and just make that canonical. I think it's a very good change for the for the technology. Another thing is we're going to see maybe a stepwise improvement there. Let's open up another bandwidth lane, okay? A secondary network that's tied and tightly coupled to the general-purpose blockchain that's just totally dedicated to a specific type of transaction. These are not necessarily new ideas, but I think the implementation we'll see in the next few months is going to be innovative for the

space because there hasn't been confidence yet so far in L2s and in general around which type of transactions we actually want to enshrine and give priority to over all transactions in the network. I think, again, something we'll probably see in maybe the next year or two is a real adoption of alternative VMs. So, you know, L1s that are competing with Ethereum have done this. This is the whole point of building a separate network and not building on the network effects of an existing one is to, you know, use a different type of technology.

And the the challenge with that is, again, you give up the economic value that's already been created through the hard work of people. And I think now that we have slowly starting to Now that we're slowly starting to see these kind of canonical trading types, it's really not um too far-fetched to think that we're going to see completely specialized blockchains. Again, Hyperliquid, going back to that, but they really like the canonical example in in the EVM space is customizing their whole execution logic for their application. So, I think this will happen over time. The other thing, again, obvious to some folks, but I think just wanted to highlight that the the willingness to put effort into these different types of paradigms really has to do with how much money you're making. I think that's something that we overlook a lot in the

blockchain space, which is revenue and building a sustainable business model. Right on the left, you have kind of things that you're trying to subsidize for growth or what I would call disruptive volume, something that you're looking to, um, to drive growth of to bring more participants into the network, but not necessarily looking to to to extract value from. Whereas on the right, significant investment, totally new VM, totally new security build-out. In that point in that model, you are likely going to see profit drivers, revenue engines, things that are going to become multi-billion dollar businesses or already are.

So, specialization is critical and I think this will continue to be a focus for layer twos, layer threes, and more. I think So, that's specialization, excuse me. The other point I wanted to go over is ordering. And so, let's talk a little bit about other information that's kind of relevant to that point. So, these are some graphs I made last, uh, two days ago using some publicly available data in in the US and, um, on Ethereum L1. So, this graph goes over, um, what is, again, my my guess at the total payment for order flow, um, money going between parties over the last 10 years using I think a combination of SEC and FINRA data. So, these are trades that are

handed between an exchange or a brokerage or maybe an exchange and another party that they know has inventory on an asset that someone wants to purchase or sell. And these are things that are handled completely out of the regular protocol, right? So, um, sometimes people call these OTC deals. There's a totally different, uh, class of them called, I think, non-alternative. I forget the last word, but this is a significant amount of value being paid for just to have access to a secondary order flow. And the reason why that exists is because if you have access to that information, like we've learned in Ethereum, you can profit off of that, right? That's how people make money in traditional finance. That's why Ethereum has looked so long to disrupt that

industry. And there's an interesting correlation here, again, somewhat conjecture, but there is, you know, a relationship here between the overall MEV value extracted on L1 and the amount of money paid for order flow in the US markets, right? You could attribute this to just general, you know, volatility in the markets, but it's also a very similar type of value exchange. An MEV boost, um, on L1, people pay for the the right to see trades coming into the network so they [snorts] can place their own trades that are relevant to that and make money off of that information. So, it's a similar action. And I think as we grow in the overall blockchain industry, you're going to see more of this payment

for order flow value transition to the blockchain. That's essentially what's already happening, but it's going to happen in a very large way in a stepwise function in my view when asset issuers move their assets and their trading books on chain as we're starting to see more parties do. Um the other thing I wanted to share show here is just, you know, the the split between these types of volume that that are happening for these OTC-like deals or again, you can see this non-ATS. Um again, the volume being processed in this non-regular exchange mode is I think 51% of all equities trades. So, that's more money being passed in a

system most people don't have access to than [snorts] you can actually access through an exchange. Again, probably common knowledge at this point, but I wanted to demonstrate it because it it shows you how much power still lies in these non-regular channels. You know, an area where if I wanted to place a trade, I [snorts] can't specify that I'd like to go talk to someone who's not advertising that they have inventory that I want access to. So, we want to move that information on chain and the parties who have the power to do so are going to come with constraints. Constraints that don't conform to the existing ordering policies, to the existing designs that a bunch of other people and in this case the Ethereum landscape

>> [snorts] >> demonstrated an ability to create on their own. So, we're going to see a shift in the rules and I think a preference for rules that are created with the help of these parties. And so, again, control over ordering policies will be paramount. I think this is going to be a huge focus for all exchanges, asset managers, brokerages, even if the clearing houses come into the mix, everybody wants the right to change how orders are processed on the network that they are having their assets transferred transferred and transacted on. So, one particular point is if you're building a general network where anything other than your transaction can take place, you're going to want those

transactions to take priority during ordering. Um this is something that I don't believe is necessarily implemented in exactly the way I'm describing today because [snorts] most of these parties don't have their exchange happen on chain. So, if it's going to happen, they're going to want to make sure that if 500 people trading another asset are activating the network at the time and they need to place a giant trail for a whale or a fund that they're working with that that transaction lands and it lands 100% of the time. So, they're going to want control over this. If you're familiar with the traditional finance structure, uh a lot of folks call this best execs or best execution.

It's just a guarantee that when I place the order and I say these parameters must be within these ranges, it will hit and they will not revert because there's a lot of downstream transactions from that. Okay. [snorts] The other thing is you could say, "No, Matt, permissionless ordering policies is the way. We have to allow these parties to become comfortable with other ways of ordering." But to them, to these parties that actually have power over the assets, right? They're compliant. They have many multi-decade or multi-hundred-year-long relationships with their constituents. That can That's [snorts] considered value leakage just to to them. They're They're saying, "I'm going to put my asset on this secondary technology, this new technology, and I'm going to give up value." That's not going to happen. That's not how business works typically. And we'll see some

flexibility as we have seen with some asset issuers. You look at stables, great example of kind of the traditional fiat currency moving on chain, >> [snorts] >> but they have rules, too. And so, there's going to be a trade-off here and if they aren't able to have these constraints, that gets in the way of the value of the blockchain, right? 24/7 trading really is the predominant value proposition for these institutions. You don't have to pay a middleman to to to settle your transaction. That's what the blockchain does. But there's going to be some preferences that they have. So, how do we balance these?

Okay. I think we're going to see basically two different types of chains that hit the [snorts] the middle mark and I think we're going to see growth of both. So, today we have permissionless systems, right? You have Ethereum, you got all these builders. And the builders are the ones who set the ordering policy for their transactions. Often times they're the ones also implementing the arbitrage strategy over those same orders. But today, you don't have to use any kind of builders, any specific one. You can go to a general network, you can make a relationship with one of them, and you have access to all of them. This is on L1.

But as we get towards the more chain-specific architecture, as these institutions all move on chain, maybe on their own, maybe with a bunch of other partners, I think you're going to see these owner-partner complex structures where you have essentially what is a committee or consortium dictate the ordering policy for the sequencer that processes transactions for their network and their assets. So, you're going to have partner transactions with priority. These are the brokerages, these are the funds that you're working with every day that, you know, essentially guarantee how, you know, the volume and and the prices of your assets flow. And [snorts] they're going to have priority on the network and everyone else is going to come through. And maybe in a in a specific faction a faction, you might see first come first serve, you might see PGA, you

might see a design like time boost on Arbitrum. But you're going to have this bimodal nature where certain parties have their application logic prioritized over everyone else and then you have kind of the rest of the market. If you don't enable that, you don't get the benefits of an open liquidity ecosystem. But [snorts] if you ignore your partners, you you kind of lose out on the existing structure. This is a separate point, but I think that in the long term we're probably going to see this, you know, to pop the stack for a minute, this collapse of all the networks with all these varying ordering policies just be completely abstracted by your your wallet. For example, you know, I want to open a levered long position on Apple and I want this much slippage. I shouldn't have to navigate to a certain RPC on a

certain purpose protocol and a certain jurisdiction to get that trade executed. Someone is willing to sell me that position. And I'm not going to, you know, compute that on my own. Maybe this is my LLM framework that I'm working with. Maybe this is my app that I'm using. Maybe this is my phone. You know, who knows exactly what it's going to be, but they're all going to spread into, you know, their own ordering policy, their own execution environment, but our transactions are going to be routed between them. And again, they're going to probably route to the best execution route for the preferences of the end user. This is [snorts] an ideal state, but just something I wanted to highlight that, you know, I focused a lot on how certain chain architectures are going to look like, but there'll be a abstraction hopefully over time.

So, what what is the point here as it relates to L2s again? Again, L2s are really, you know, a term used to describe a specific roll-up that's settling through Ethereum, but you can build chains that post any arbitrary data to any chain in EVM and outside of that. The point here is that sovereignty over your chain is what matters for these institutions. So, L2s, L3s, L whatever you name it, that's what works for them. Sequencing can be handled by one party or a fixed committee. And what this enables is the changing of those ordering policies to happen way more quickly, right? You don't have to find some arbitrary builder or coordinate with the whole network to agree on a hard fork. You can just update that with a vote. The other thing is, as I mentioned, custom

execution logic. It's going to change over time. You know, the way that we do transactions today is not enshrined. And so, as more things become valuable and dominant, you can add those to the the execution logic of the chain again on your own velocity. This is the whole reason why we built L2s in the first place, right? To separate the execution from the settlement. And then I think really importantly, cross-chain actions are really really only possible in a good way for the user by having sharded networks. And so, the way we get there, I think some others were talking about base roll-ups, but in our minds, using the the technology to have a separate network and configuring that to your needs is the right way to do it. And I think you can argue on this

on the security the security angle, but ultimately flexibility is a choice, not a compromise. It's up to the party to just to set their own profile for what they're willing to put on the table. Okay, so these are some other random points in the last couple of minutes uh that are coming to L2s. You see optimistic and ZK talked about a a lot about separately, somewhat of an old narrative at this point, but just to, you know, refresh folks, they are different architectures. We believe that there will be a hybrid approach where, you know, at some point you'll have real-time ZK proving. The hardware will obviously logarithmically become cheap enough to use. But in the moment, the soft finality guarantees you get from using a sequencer and knowing that the ordering is deterministic um in in EVM is good enough. So, I can

transact across chains relying on that guarantee and that's my relationship with the party that's facilitating that transaction. I think native cross-chain balances, which is something that account abstraction frameworks like ZeroDev support, is going to enable users to spend assets on any chain. This is very very important to to have tra- to have DeFi feel like a bank. And I think something that is an interesting concept that we we hear a lot of work on agents. I think very good standards coming out in the space that dictate how to transact on chain, how to interact with different applications.

We have I have not personally seen them spawn their own chains, but this is something that can be done. This notion of agent-to-agent commerce creating their own network for a small amount of time to resolve a large amount of transactions is something that is technically possible and it's up to us to facilitate that development. So again, protocols that focus on transitioning new business on chain and supporting real value creation will win. I think I've heard one of my coworkers say this, it's revenue season, baby. Like it it's true. Building a sustainable business is a actual thing that a lot of protocol teams can ignore and I think particularly when we're building technology now, we have to confront that reality.

So, I like to end my talks with just some open questions to consider as you as you go out on your day. One thing is I'm wondering if we'll start to see programmatic enforcement of how stables can be used on chain. We see rules on how they can be mint and burned, but there's no control over the action. Interesting. Um And another thing, will we see a resurgence of functioning DAOs that map to real life structure? So, the Arbitrum DAO is a great example, something new, something that works, but it's only specific to one network. DAOs in general have not really been uptaken by anyone because they're incredibly hard to manage. Collective action is, you know, one of the prevailing problems with general policy.

But, when real money starts to be managed on chain, I think then we'll have a real impetus for the parties that are conducting these transactions to take advantage of the cryptographic guarantees they have around voting, around making decisions together. And that's when you see the value that folks were talking about, you know, 5, 6, 7, 8 years ago actually start to become a reality. Cool. Thanks for listening to my talk and if we have a couple minutes, I'm happy to take questions. Otherwise, have a great day. No time.

Awesome. Thank you. >> [laughter] >> All right.