Launching a rollup is the easy part

Transcript

Test test test test. Sweet. All right, I just got in from like a 30-hour travel days and I did these slides during that. So, if this talk sucks, that's probably why. Uh all right, today we're going to talk about launching a roll-up and why that is actually the easy part. Um so, Kevin and I work at Optimism and for today you all are going to work at a fintech. So, think of something like a uh like a Revolut or a Plaid or a Venmo or a SoFi or some sort of a uh neo bank or like tech forward bank. Uh and so, the hardest thing I'm going to ask you to do is to remember what it felt like to not

know anything about Ethereum or DeFi or L2s or or anything like that. Uh and so, this is crypto for you. Um it's it's a bunch of coins, it's a lot of scams. Um you don't really know much about Ethereum other than you kind of wish you sold it at this point. Um but to help you get into this role, uh I had the opportunity to interview about 10 uh engineers across uh tradfi, fintech, and uh web3.

Um and just first question in each uh interview was essentially rate your crypto knowledge from one to five and on average uh the tradfi and fintech developers rated their knowledge higher than the uh web3 crypto developers. So, to give you kind of a a sense, uh you are a confident uh crypto expert uh and you know everything that you need to know about crypto. So, that brings us to our first problem, uh which is the general knowledge gap. So, your understanding of of crypto is that everything is like Bitcoin, everything's competing with the dollar.

Uh you don't really understand that the real use case uh of crypto is new financial rails and you don't understand that stable coins are the first killer use case of crypto uh and that like the dollar uh is solving a lot and not necessarily uh attempting to be replaced. Um you're constantly seeing headlines about scams uh and you've you've read about DeFi, but you don't really understand that uh it's it's going to replace the back end of finance, uh not necessarily the front end.

>> [snorts] >> Um and you don't really understand that in the long term the average user uh won't even need to really know that their money is moving on chain. Uh but luckily for you, you know a guy. Uh you've had an engineer on your team who hasn't shut up about Ethereum since like 2017. Uh and you finally give him the time of day uh to have him explain uh about Ethereum uh and he tells you that replacing your tradfi rails with DeFi will save you a bunch of money. Uh so, he starts sending you headlines about how finance is moving on chain, whatever that means. Uh he shows you a bunch of headlines about a bunch of really big tradfi banks that are moving

on chain. Uh [snorts] that the New York Stock Exchange is developing on chain settlement. That uh tokenized stocks like gold are moving on chain. Uh that a lot of your competitors in fintech have already moved on chain. And that a lot of CeFi companies uh are building their own chains um and expanding their business models to incorporate DeFi. Um and so, realistically all of finance is moving on chain. Uh and that's because DeFi as your uh nerdy little Ethereum bro tells you uh DeFi is an order of magnitude improvement over tradfi. Um the

attributes of DeFi are far more suitable to the needs of uh modern financial products than tradfi. And the attributes of Ethereum are far more suitable for DeFi than any other L1. Uh and he tells you that DeFi is open, public borderless neutral censorship resistant permissionless disintermediated disintermediated. Transparent fast cheap global peer-to-peer, non-custodial, and easily innovated. And that tradfi is none of those things. Uh and he he gets all excited about all these other use cases like reducing platform risks and aligning incentives.

Uh increasing trust in institutions, unbundling the layers of social media like separating the social social graph from the algorithm from the content. [snorts] Um and overall just reducing the need to give billionaires complete control over the internet. Um personally, you don't care about any of this stuff. You're just trying to make that money for your company. Uh but regardless, you're interested and it seems like this is like an interesting way to do it. So, you put together a strike team and you start learning to develop software for Ethereum. Which brings us to problem number two, the technical knowledge gap.

So, learning how to develop software for the EVM is a lot more than just picking up a language or framework. Um the code literally carries money and costs thousands of dollars to deploy it. Error messages [snorts] are either encoded or just not existent. Um and overall web developers coming from a web2 background um it's very unintuitive uh to their use of prioritizing speed above all else, which is not an intuition that scales or that translates to L1 scaling.

But your uh Ethereum engineer explains layer ones and layer twos and ECDSA elliptic curves and uh contracts and wallets and and the roll-up centric road map. Uh but the this complexity just introduces a ton of unknowns. Uh and you start to slip on your timelines. So, we get to uh problem number three, which is the regulatory gap. It's a very unsolved gap in most of the world. Compliance is hard and complicated. Um and a lot of times this really just looks like risk minimization, uh risk-reward optimization than it is really like following specific guidance.

Um but your Ethereum engineer tells your legal team that running a roll-up is like you can kind of control some stuff and then other stuff you don't really control at all. Um and then he teaches your developers about sequencers and escape hatches and blobs, proto-danksharding. >> [snorts] >> Uh and there's even more unknowns um and your CEO is starting to ask why the project's taking so long. So, you tell him about how real-world assets have expanded to uh over 20 billion already on chain. Uh [snorts] and up to 300 billion if you include stable coins. And he's interested. He's he's still on board.

Um and that brings us to problem number four. So, research and development. He realizes a constantly growing, evolving ecosystem. New EIPs are written daily. Uh new hard forks roll out every six months. Uh it's pretty clear that you're going to have to uh have a serious ongoing investment if if you're going to have an L2. And so, team is, you know, starts out writing optimistic roll-ups with fault proofs and then switches to ZK roll-ups with validity proofs cuz apparently those are cooler. Um [snorts] and so, just again, further milestones are missed.

Um [snorts] your boss is angry. You explain that sub-cent instant global transactions of any size are now standard. Uh that stable coins processed more value last year, 27 trillion, than Visa's entire annual payments volume. Uh and that JP Morgan moves 2 billion daily on Ethereum rails. And uh the CEO has no idea that this much adoption's already happened behind the scenes. So, it brings us to problem number five, which is stack choice. Um your engineers finally ship some layer two smart contracts. Uh that's the good news. The bad news is that it failed the audit and so, you frantically Google the most popular open-source Ethereum L2 stack, which brings you to Optimism.

Um back in November the sum of fees generated by dApps in the OP stack chains grew past five 852 million. >> [snorts] >> Um just a couple random facts. OP stack chains earn up to 3x more profit uh per dollar on chain than the most successful chains on other L2 stacks. It's has 50 plus chains in production, 6.1 billion TVL, 58% of all L2 transactions are on the OP stack. Uh and it encompasses 14% of all transactions in crypto. So, we get to problem number six, which is partnerships.

Uh you've finally uh rolled a OP stack L2. Uh and your Ethereum engineer tells you that that's just the easy part and that the next thing you need to do is figure out a bunch of peripherals. So, you're going to need nodes, RPC providers, sequencer, block explorers, bridges, assets, external wallet support, embedded wallet support, transaction simulators, oracles indexers DEXes lending protocols, yield aggregators, market curators, liquidators, risk managers, monitoring, compliance. Uh and for all these, there's like half a dozen options. And your legal team is going to have to draft and review bespoke

financial contracts with every single one. Uh and your CEO is not happy. Uh so, you show him this case study of Coinbase. Um And the way that they launched an L2 and then continued to launch products and services on chain to integrate uh deeper into DeFi uh as their TVL continues to grow. A third of Basis TVL today comes from their C 5 integrations and it's expected to contribute to over 50% in under a year. And then in the last year they scaled to about 2 billion in deposits earning their users 4 to 6% on stable coins along the way, which is obviously much higher than any trad by bank.

And these are hyper scalable services. So incremental [clears throat] complexity of additional markets is zero, cost of capital to offer on chain loans is zero, balance sheet impact is zero, cost to maintain licensing is zero, origin of origination fees is zero, and it's cheaper access to credit than credit cards. So you show them that their on chain borrow product continues to grow up into the right and your CEO chooses to let you to continue the project. So next problem is the economic cost. It's you're coming to the realization that if [snorts] you combine the the engineers that you needed to hire with EVM skills, the dozen enterprise partnerships and legal fees

with the infrastructure cost alone, which can be half a million to a million or multi-million depending on usage, running an L2 is just incredibly expensive and especially up front. And so you learn about OP Enterprise, which offers a full spectrum of services depending on your size and use case. So on the first OP main net, which lets you start [snorts] with a a proven competitor agnostic public network and then graduate to your own chain when you're ready. Then self-manage where you operate the roll up and then we support you. And finally fully managed where we take care of the entire chain so that

you don't have to. And in the 11th hour your engineer decides to pull up Actions SDK where he claims he can recreate the modern fintech DFI playbook with it. It's an open source typescript SDK that lets your users perform on chain actions like lend, borrow, swap and pay. It supports embedded wallet providers like Privy, Turnkey and Dynamic, which make gives your users access to wallets tied to their existing email off accounts. Lets you lets you create smart contract wallets, manage permissions and configure N of M custody models that work for with your legal team.

It supports lend providers like Morpho, Aave. So your users can earn yield without even realizing they're on chain. It supports swap providers like Uniswap and Velodrome. So your users can trade between allowed listed assets and markets. It lets you import the most popular assets so you don't need to hard code a dozen contracted addresses across a dozen chains. And it also lets you add your own custom assets. And it additionally supports Ethereum main net, OP main net and all OP stack chains. So your users can seamlessly move between assets and markets across chains without them ever needing to know about bridging.

Um And so your Ethereum engineer creates a feature branch which lets your users click a few buttons and move their USD from a trad by checking account earning 0.1 interest into USDC earning 6% interest in DFI. Um it lets you determine which control which wallets control which funds along the way. Choose which DFI protocols you want to support, which markets, assets and chains along the way. And so you launch your new DFI integration.

>> [snorts] >> The initial use case of integrating DFI from an existing chain to proves the viability that moving your business on chain actually makes sense. And then over the next quarter you can work with the OP Enterprise team to figure out if and when an L2 makes sense for your business. So some additional takeaways from this. The two over simplified reasons why companies are moving on chain right now. One, introducing new product lines, so like Coinbase's earn and borrow products.

And then two, saving costs, simplifying processes, automating, avoiding international remittance fees, etc. Start small. So if you're a C 5 exchange or a fintech or some sort of a neo bank, you don't need to start by rolling your own chain. You can just dip your toes in and integrate DFI on existing chains. And then we obviously still need L2s. Vitalik wrote a blog post a little while ago and people misconstrued it as him saying that the roll up centric road map is dead. It's not what he's saying. You should read it. Um It's like saying that you don't need horizontal scaling or that like

vertical scaling makes horizontal scaling obsolete. And it's just like that's not really how any of this works. Um I think I mean if there's one big takeaway here it's that L2s are basically 20 enterprise partnerships and legal agreements under a trench coat. And that the contracts are an implementation detail. And if we want to get to a world where we bring the world on chain, we are going to need to make it as easy as provisioning cloud infrastructure.

And OP Enterprise is our particular investment into that future. So we'll help you navigate the problem space, make partnerships and even help you run the infrastructure. But it's a problem that needs to be solved at every level of the stack. So I hope other teams join us. Um and then having a composability plan once you do have an L2. So figuring out how your L2 fits in with the rest of the ecosystem, what you can leverage and then what you can kind of offer back. Every open source solution across the OP stack goes back to benefit the collective. Others chains can use it. Um if you are curious to hear a little bit more about OP Enterprise, reach out to us. And then you can follow me on

Twitter at Kevin_Weaver. That's my Twitter QR code. I'm working on Actions SDK. So if you want to talk about that, come check it out. And that's it. >> [applause] >> Any questions? Nice. I'm going to go sleep. Thank you. Thank you. Oh yeah. like the art of people and how they should build it. What happens if someone makes a choice and it's like the wrong choice? Is that going to make it a more painful process?