Fireside Chat: Arbitrum, L2s and Institutions with Steven Goldfeder and Amanda Cassatt

Transcript

[applause] >> I'm super excited for this conversation. Um to get started So, I think you literally co-wrote the textbook that a lot of this room learned crypto from. Um and Arbitrum started as your Princeton research project. Take me back, did you have any idea what it would become? >> Absolutely not. Um actually we discussed in the earlier today we were in the in the back room what did I want to be when I get older and it actually was a professor for the longest time and uh I actually remember doing an internship at Google once and there was this other uh friend of mine and we would argue all day long. I'm like,

"Why would you ever like build something? I mean, you should be a professor." And he would say, "Why would you ever like go into academia? You should build something." Today he's a professor at the Hebrew University and I have founded a company, but uh no I uh never imagined where this would go and actually if you look back at the very very first mention of Arbitrum on YouTube, you'll find a fall 2014 Princeton class project which actually was my co-founder Ed leading this project before I was even involved in the project at all. And um it was very much an academic pursuit. The reason I like to mention that is Ethereum as you know went live in July of 2015. This was a class that took place the previous year, so almost a year earlier and it was really an academic pursuit. The idea was, "Hey,

there's this theorized thing called smart contracts. Um and if anyone ever gets interested in them and actually wants to use them, which looks like they might cuz they're pretty powerful, they're not going to work at scale and we need to build scaling solutions." And it was very much an academic pursuit and it was really only uh years later that um when it came time, you know, I gradu- I was graduating we actually had smart contract platforms. They actually started running into the initial scaling issues that I said and so did my co-founders at an Harry, hey, let's take this out of you know the academic context and actually build this and here we are 8 years later still building building this vision.

>> Out of the ivory tower and into the streets. >> Exactly. >> Um so fast forward to the very proximate future. Vitalik just said that the roll-up centric roadmap direct quote no longer makes sense and that we need a new path. The founder of the chain you've built on seems to be questioning the whole category. Is he right? He warned that an L2 that doesn't fully decentralize is really just a vampiric L1 with a bridge. What makes Arbitrum different from the chains he's worried about?

>> So I think there's a few things that that happened here. Number one is there are a lot of different projects that call themselves layer twos with very different security you know characteristics. And for many years Ethereum took a very neutral approach where hey, we're not going to call out anyone, not going to do any do anything and like basically everyone can use the Ethereum brand and it could be whatever they want to be. Clearly that doesn't work because you have chains that are not really connected to Ethereum, aren't secured by Ethereum, but they've been latching onto the Ethereum brand. So I'm actually very happy with that ending. Arbitrum is different, you know, from day one.

Arbitrum is unique that it was built with the proofs that connected to Ethereum that get its security from Ethereum that leverage Ethereum's decentralization. At the same time there definitely has been an evolution between the time where Vitalik initially presented the roll-up centric roadmap to today. And I would say there are two core things that have changed. Number one, back then the initial users of Arbitrum say in 2021 were Ethereum users that wanted to use Ethereum cheaper and that was Arbitrum's entire identity. It's hey, I want I like Ethereum, I like the EVM, I like its security, I just want a cheaper path, so therefore I'm going to bridge to Arbitrum. But, if you look at Arbitrum today, there are still those users, but I would say the majority of users are actually the Arbitrum first

users. Users that say, "Hey, I like Arbitrum's deep liquidity. Hey, I want to use USDAI or Variational or Osteum or GMX or one of the many thousands of Arbitrum applications, and that's what's driving them to Arbitrum. Hey, I like Arbitrum's 250 millisecond block times." Right? So, there are the Arbitrum features that are actually driving them there. I think that's one large thing that has changed where Arbitrum has built its own identity. It's true that it's Ethereum's number one layer two, but it's also just a top five blockchain by economic activity. And the second one is scaling actually developed an identity I sorry, L2s developed an identity that's actually somewhat independent from scaling or has a path that's independent from scaling. What I mean by that is the following. Imagine that you could

infinitely scale Ethereum layer one to the point where you didn't need L2s for scaling. We're far from there today. There are benefit many benefits of having a two-layered approach, but just imagine for a second that we could infinitely scale layer one. So, you would say, "Is there still a need to have Ethereum layer twos?" Back then, I probably would have said, "No." But, today I would say, "Yes." Because if you look at what's happening in the institutional space, it many institutions come from the starting point like table stakes is we want to build our own chain. Not because they don't think that Ethereum or Solana or whatever other other layer they might use can scale. It's cuz, "Hey, we want to use build our own chain." That's just because we want our own branding, we want our own compliance requirements, we want our own privacy. They have a host

of reasons and many, you know, all valid that hey, they want their own chain. And what are institutions doing today? They're choosing between between either building their own layer one or building an Ethereum layer two. Right? If you take the four foremost prominent institutions that have announced their own chain in the past year or so, so it's Stripe building Tempo layer one, Circle building Arclayer one, Robinhood building on Arbitrum layer two, and Coinbase of course building base a layer two. And that's the interesting thing that's happening today. And if you pull back layer twos and say, "Hey, we don't want to build layer twos anymore."

they're not going to build any Ethereum. They want their own chain. They're just going to go the way of Stripe and Tempo and Circle Arc and build layer one. So, interestingly enough, in the institutional battle that exists today, if you accept the fact that institutions, the largest institutions want their own playground, want their own chain, Ethereum is is the only layer one that actually can fight and say, "Hey, we can actually cater to you and give you that via Arbitrum, via layer twos, via the roll-up centric roadmap, even if it's completely separate from scaling." And hey, I believe in scaling the layer one.

We at Offchain not only build Arbitrum, we build Prism, which is the leading consensus client. And every major layer one upgrade that's happened from the merge, you know, all the way up to to now, we've been a very, very big part of cuz we believe in scaling the layer one, but that's not uh at all in contradiction to to having layer twos both for scaling, for features like Stylus, for features like fast block times, but also just for institutions that want their own chain. >> So, you would say that in this world where um L1 fees have collapsed and it's become quite uh inexpensive to use the Ethereum blockchain, which used to be the value prop of L2s, that the case for L2s today is really institutional

onboarding? >> I think it's uh two things. Number one is scaling goes like this. So, I don't think that we're, you know, past the era of expensive layer one block space. I just think that uh you know, it ebbs and it flows. So, we have to always be prepared not just to service the needs of today, but how do we service the needs of the next 10 years? You see every large institution coming on chain. The fact is one blockchain cannot support them. And that's the beauty of of uh the two-layer approach. Uh but yes, for The other thing is for institutions, even separate from a scaling, so even if you believe that L1s would be able to service everyone, they want, you know, their own custom playground for many reasons. Some of

it's just branding. We just want to say, "Hey, this is, you know, base or this is Robinhood chain." And that's fine, you know, we want our own branded space. Uh there's also another benefit of actually collecting fees and saying, "Hey, you know, if I'm a large application that's driving millions and millions of dollars of fee revenue, I'd rather actually own the chain and collect that as revenue rather than have that as an operating expense." Really, really logical uh you know, decision for particularly for a large institution that controls a lot of liquid a lot of liquidity and also controls retail users and, you know, a Robinhood user is not coming to Arbitrum because they're like, "Hey, we need to be in Arbitrum." They're coming to Arbitrum because the Robinhood app is sending them to Arbitrum. If you're in that position,

you might as well say, "Well, if I'm actually controlling where where these users are going, I might as well uh profit from that." You know, in Vlad Vlad from Robinhood's words, in Vlad's words, he says, "I want to be a landlord and not a tenant." If you imagine you have like a strip mall and you have like a bakery that's like super, super, super popular and everyone's like, you know, lining up to be part of your bakery, the convenience store next door is also going to do really, really, you know, good business because it's, you know, co-located in close proximity to your bakery. Now, if you own the strip mall, you're going to actually benefit from that. If you don't and you're both just tenants there, then all that ancillary activity that you're driving actually is not going to uh you know, um really flow

to you at all. So, there are a lot of valid reasons why institutions want their own chains, compliance, privacy, governance, and I think that in addition to scaling, I wouldn't say scaling is is is not a valid reason anymore, but there's a whole another reason that's completely independent from scaling. >> So, let's say um he Vlad and um and Robinhood, let's say he's the landlord. What what's Arbitrum in that case? [clears throat] Um I guess my question is if you're sort of the a AWS of blockchains, if you're inside and no one actually sees Arbitrum under the hood, you know, right now when you buy a stock um on Robinhood in Europe, it's actually a token um on a an Arbitrum uh orbit chain, right? And so

um are there risks to being invisible infrastructure? And where where are you in his metaphor? >> There are definitely risks, but I actually think it's the absolute best place to be, particularly if we're focused on global adoption. If we're happy with the, you know, number of users we have today in crypto, then and the users that are mostly enthusiasts that actually understand what Arbitrum is and understand what a wallet is and what a seed phrase is and are comfortable using this technology, then sure, I don't need to fade into the background. But if we zoom out and say, "Wait a second, the number of crypto users today is actually tiny compared compared to the global financial market." Like how do we grow that? The answer is you're going to need to grow that among a class of users that don't really care about the details of the

blockchain. They just want the experience and the benefits that the blockchain can get them. Robinhood users in Europe being a great great example. So, Arbitrum is fueling their international expansion uh today on Arbitrum one, eventually on their own on their own orbit chain. And like you said, you go in the app, you're getting a tokenized, you know, you want to buy Apple, you're getting a tokenized version of Apple, but like you don't even need to know that or pay attention to that. You just say, "Hey, I want exposure to Apple." You go through the same exact flow, you know, modular modular little tokenized uh uh pill that they have there as you do in the US app and you get this uh yeah, and that's actually I think is a big success for infrastructure because it means that we're now able to cater to, you know, the people who don't know how

to back up their seed phrase, who don't know what a wallet is, who don't want to know what a wallet is, who don't want to uh go online and think about the network. An example I like to give is imagine like if you went on YouTube to watch a cat video and this uh pop-up came up and it said, "Would you like to watch this on AWS today or on GCP? Right, on Amazon servers or Google servers?" And you're like, "I don't know what any of that means. I just really want to watch a cat video." Right? So, we need to basically abstract this away from users. And sure, that means we're going to go into the background, um but also in Arbitrum, the economics are set up that Arbitrum, for example, in Robinhood's case, the Arbitrum DAO, the token holder controlled wallet, gets 10%

of the revenue from that chain. So, the economics are such that even if you're in the background, you know, the ecosystem is still benefiting uh in dollars and cents in real ways from that traffic. And ultimately, the only way to reach global traffic is going to be uh from fading into the background a bit. >> So, speaking of the user's experience who's maybe trying to just watch that cat video, um right now on Ethereum, we see this fragmentation amongst all the L2s and L3s. Um And bridging is still a nightmare. I guess my question is as a champion of L3s um and Orbit, um what's our what's our path back to feeling like one Ethereum?

>> That's today, I would say, the biggest open frontier for development. And it's one that we at Offchain are actually embarking on very strongly. So, I mentioned we built Arbitrum, of course. We build uh Prism, I mentioned, which is the Ethereum client. And we also built something called ZeroDev, which is a the leading smart wallet. And one of the things that we do is uh bridging and uh chain abstraction, right? So, bridging without feeling like you're bridging. We have a product called uh smart routing addresses, which basically means if you have a DeFi app, say on Arbitrum, and you want to extract liquidity from anywhere.

Um most users will actually fall If you have, say, you know, dollars on Ethereum or dollars on some other network, and you have to Maybe it's USDT there, it's USDC here. Maybe the gas fee there is in ETH, but in the other network, like it just it's a crazy, crazy experience. And you know, many, many steps for users just to bridge over their funds from one network to another. And to just, you know, uh enter a DeFi app. So, this uh smart routing address product says, "You have an app on Arbitrum, but we'll give you an address in any chain, right? So, Ethereum, Base, Solana, you have an address in that chain." All users need to do is send their funds to that address, and all that bridging happens underneath the hood. I think this is the frontier and the biggest problem that we need to solve today. And you might say,

"Okay, but like, why did we create that problem?" And the answer is sometimes you need to, you know, take a slight step backwards in order to go forward. Um and actually a good analogy in this is building a company, right? So, uh you know, we have almost 200 people in our company today, and the way we operate the company today is very different than we did when it was 10 people. When it was 10 people, I would go into a room every day and say to every single person, literally, "What are you doing today? What are you doing today?" I would know what every employee is doing. Eventually, we had to break that up, create other structures. And in the initial outset, that was a step backwards, right? I was less efficient on day one, but we would never get to where we are today in the amount, you

know, with the with the initial structure. And that's I think the same thing with chains. Yes, one monolithic chain works much, much, much better for the traffic it can support up until this point, but then everything starts breaking. And if you don't actually take the benefit to invest in infrastructure, you actually will never get past that point. So, you have to say, "Okay, we're going to, you know, separate this. We're going to build a layered approach. It's going to be a little painful in the beginning, but ultimately, it's the only thing that's going to enable us to support global commerce." >> So, one of the solutions that people always bring up to this fragmentation question is, "Oh, the AI agents are going to solve it because they themselves exist at a layer of abstraction above all this mess where

they'll they'll just solve all of it, and we'll each just be talking to our agents, and they'll, you know, deal with the whole thing." Is that your view? I mean, how are you preparing, if you're preparing, Arbitrum to interact with agents? Do you imagine that agents are going to be the majority of the users on Arbitrum? >> So, I don't subscribe to the view that agents are going to solve this for us. In fact, I subscribe to the view that agents are going to add to the you know, usage and the bandwidth constraints of the network. Because I do think that there will be a lot of agentic commerce. We're already beginning to see this. We're seeing protocols like XYO 2 that are directly targeting this.

And we are supporting those and you know, I think Arbitrum will be a very fundamental and key layer for agents to interact. But again, I view that as that's not going to make less human interaction. That's going to add more usage to the network. Potentially, if you don't build it correctly, you know, add more congestion and and and and add more constraints to this problem that we solve. So it becomes even more important to solve. Um >> Like AI is a giant civil attack, basically. >> It's just many many more users. Exactly. Yes. And with you know, they can do microtransactions and just add a lot of traffic to the network. And I think it's fantastic. And I think crypto is the right set of tools for many many AI applications and giving an agent agents

wallets with particularly smart wallets with policies and constraints like we do in Zero Dev that have like smart smart contracts that actually enforce policies. Or also trusted hardware that actually enforces policies on wallets, which you'll need cuz you don't want to give your agent unfettered access, obviously, to your funds. So you can have a really nice interaction between crypto and AI. But yes, I think that it's closer to a civil attack than a solution, for sure. >> When we imagine Arbitrum and Ethereum, you know, 5 years, 10 years out into the future, do you think that people are going to know that they're using Arbitrum or know that they're using Ethereum? Do you think that we'll look back and think it was crazy that we made cults around back office software? Or do

you think these things will still be vibrant I know no better word than to say but brands. >> So, I think the a lot of the power of crypto in general, right? So, Ethereum, Arbitrum, but just the ethos of decentralization and self-sovereignty of your assets is that these tools exist. So, you know that you're not getting locked in, but the truth is, particularly as we grow to a global user base, that many users actually don't want to, in good times, you know, take advantage of that. What I mean by that is, you know, many users would prefer to have a bank hold their assets than keep their assets under their pillow or in their ledger. And that's just like a fact. And for many users, that's actually probably the better uh the better approach given the security

constraints of both. Nevertheless, the fact that it is an open economy, the fact that there are these permissionless escape routes, if you will, these these escape hatches that say, "Hey, if you don't like where you are, ultimately they are your assets and you can take control of them." That's extremely powerful. And I think we need to build a uh an ecosystem that understands and leverages both those users. I think gaming is is a very good example here. You know, in the early early days of blockchain gaming, it was the Ethereum enthusiasts saying, "Hey, oh my oh my god, I can actually play a game on Ethereum. That is so cool." But the game didn't actually need to be very interesting because you were just basically uh showing off to those that liked Ethereum that, "Hey, look what Ethereum can do." In the second

generation, it was just, "Let's build good games as the first principle, but also say, 'Hey, where can blockchains help us there?'" And this is where I think Robinhood, you know, the example there is so strong because, you know, Robinhood is um basically saying, "Hey, users, you can actually get the benefits of the blockchain. You can join the programmable economy, where we can actually take these markets, launch them on the blockchain. You get all the benefits of the automation of this market, but you don't actually have to even engage about engage with it engage with it or even know that you're that you're doing it." So, to answer your question, I think most users won't engage with that, but they will get the benefits of it, right? They'll have, you know, the benefits of the programmable economy, the benefits of markets that are 24/7, the benefits of instant

stablecoin settlement that are just much, much better and massive upgrades the state of finance today. But they probably won't engage in the details the same way when you when you go to YouTube, you don't really think about I wonder where this infrastructure is hosted. But there will be a class of users, particularly when the you know, the system is tested and there's like, you know, attacks in the system where people feel uncomfortable. Like we see this like FTX is a good example, right? Like, oh, maybe decentralization is important. There will be users that are consoled by the fact that hey, I can take self-sovereignty of my assets. I'm not locked up in the system. Even if I actually never exercise the right, the very fact that that right exists, I think is important and always will be in the back of the mind of of many users,

even larger than the number of users that actually utilize it. >> Just because it's not raining right now, it doesn't mean that no one should own an umbrella. >> For sure. >> Um I'm so glad you brought up gaming um because right now in our industry, it feels to me like people think only the financial use cases will or do have product market fit. RWAs, stablecoins, um on-chain credit, um DeFi, those kinds of things, uh but not anything else. Are you bullish on any non-financial use cases?

>> So Arbitrum has been very finance-focused even before actually that was cool. If you look at where perps took off, right? Now today obviously perps are one of the most exciting uh use cases and the most important use cases in crypto, but like, you know, in 2021, there was only one blockchain that was doing perps, that was Arbitrum with GMX, with Gains Network, and really perps took off in Arbitrum. So we've been very uh financially focused from the very beginning. Uh that being said, I do think that there is a large class of um applications that uh the blockchain can help and will help and often there are financialized components of them, but they're not just like pure everyday finance or their expansions of what you know, previously we had done. Prediction markets are a great idea. Obviously,

they're markets and they're very, very financial, but they're an expansion of, you know, the typical market structure that we've previously used. Another good example on Arbitrum is USDAI. So, this is again solving a world real-world problem. And this is the problem of if you want to purchase AI hardware and get funding for it, there's like a middle market problem. If you're like, say, between 30 million and 100 million dollars, which is a lot of money, but you're not enough to be like one of the really, really big players trying to procure hardware, it's very hard to get funding and there's a protocol on Arbitrum that's, you know, backed as a that's presented as a as a stablecoin that's actually providing lending for this middle-market problem. So, it's financialized, but it's actually

funding entrepreneurs building AI in this middle-size market that doesn't exist anywhere else. So, I do think that, you know, Arbitrum is the finance-centric blockchain programming powering the programmable economy. But what is the programmable economy? It's the idea that you can put markets on chain and those markets are often very varied and also expansive and the definition of what markets can be are much larger than they were traditionally. So, I think there is a finance edge to most, if not all, applications today that are successful on the blockchain, but like it's definitely not necessarily the only part of it and it often there's a real-world aspect that's much larger than that.

>> Last super quick question. I think something that a lot of community really treasures about Ethereum is that it does have a specific mission and set of values that are maybe orthogonal to the business case. Long like crops, right? Censorship resistance open free private permissionless secure. How does the Arbitrum mission and business fit into the set of Ethereum values? >> So, I believe very strongly in in those values and in fact, that's what brought me to the blockchain in the first place, right? Like I didn't uh you know, devote my career to building Arbitrum and building the Ethereum space to replace one centralized system with another. You

know, you can make a lot of money off it. Some the the new uh power brokers in the new centralized system will make a lot of money, but ultimately that's not what we're doing here. And then I think it's very refreshing to see, you know, Vitalik uh so committed to uh you know, the philosophical uh reasons that brought him here. And uh even when those are unpopular, that being said, I also think it's important that we see enterprises and the biggest industrial players wanting to build on Ethereum today. And we need to have an avenue to them, not just a philosophical theoretical partner, but an avenue to say, "Hey, we can support you. We can do this." And I think that what we're moving on in the Ethereum ecosystem is just a broader set of companies that are operating here.

So, the Ethereum Foundation is going to be very, very focused on crops, and that's great. And you're going to have companies like Offchain that are going to be the the business conduits to say, "Hey, institution, we can connect your assets and your aspirations to actually build on chain to actually build on Ethereum to bridge the gap between the philosophical crops and whatever products you have today. And I view this as a very uh harmonious ecosystem where uh companies like Offchain and the Arbitrum ecosystem are going to have to step up to fill uh a void, but it's a void that I think we're very, very uh well positioned to fix and also dovetails well with the more philosophical approach of the Ethereum Foundation.

>> Well, building a vibrant farm uh in the infinite garden. So, thank you so much. I think that's a great place to wrap it. >> Thank you.