Public Chains, Private Guarantees: Designing Layer 2s for Institutional Settlement

Transcript

Okay. Hello. Uh thanks everyone. Um so, I'm Declan Fox. I run Linea at ConsenSys. Uh for those who don't know, ConsenSys is also the company behind MetaMask, uh Infura, and Hyperledger Besu. Um we've been building Ethereum infrastructure for over a decade. Uh so, today I want to talk about the most important question facing Ethereum right now. Can we capture institutional finance? You know, not in theory, but in production with real money. So, the title of this talk Public Chains, Private Guarantees, that tension is exactly where the opportunity lives.

Okay, so let's start with what's actually happening. Uh Ethereum settled $14.3 trillion in stable coins last year, uh which is more than MasterCard's $7.8 trillion. BlackRock Biddle hit $2.9 billion in tokenized treasuries across nine chains. Total tokenized real-world assets are somewhere between 19 and 24 billion, up 380% from 2022. And 60% of the Fortune 500 now have active blockchain initiatives. So, the debate about whether institutions come on chain, that's over.

The only question left is which infrastructure do they choose? And that answer is going to determine the value of every asset in this ecosystem for the next decade. Okay, why is this happening now? Uh four things are all happening at once. So, we have first regula- regula- regulation finally crystallized. Uh MiCA is enforced across EU. Uh we have the Genius Act signed in the US. So, for the first time institutions have a clear legal path for doing things on chain.

Second, the permission chain story is falling apart. You have R3 which pivoted to Corda, uh Solana. Uh IBM abandoned Fabric. Uh the vendors who promised private chains for enterprises have all either pivoted or are dying. Uh third, institutions want settlement that works around the clock. T+0 atomic finality, repo, delivery versus payment, intraday collateral, all atomic, all programmable. And fourth, the enterprise L2 stack has caught up.

ZK proving, zero-knowledge proving, works in production right now. Privacy and compliance are built natively into L2 infrastructure. For the first time, public chains can actually meet what institutions need. Okay, uh this is a real quote uh from one of the world's largest custodians that said told us this directly. Um they want to replace 60 to 70 internal systems with a single on-chain ledger per asset class. And I think this quote really captures the scale of what we're talking about here. You know, this isn't just about tokenizing a bond so you can put out a press release. This is about replacing the actual core infrastructure of capital markets.

And for 5 years, you know, enterprises were told, you know, privacy or Ethereum. Uh pick one, okay? And honestly, that was true. Um the enterprise tooling wasn't there, so institutions, they would build on private chains. Um Canton exists because of that gap. But the architecture has caught up, and now the choice is different. So, there are really three visions competing uh for this. Um Canton is somewhat ahead in terms of uh incumbency. DTCC co- co-chairs their foundation. Goldman Goldman Sachs and Nasdaq are members. Uh they're doing $280 billion daily in US Treasury repo.

And you know, honestly, their their privacy model is pretty strong. Um but you do pay for it, right? They have Daml lock-in, no Ethereum settlement, no access to public liquidity. Um And then you have Solana. Okay, so Solana is fast. Um Visa chose them. R3 is bridging Corda assets over to Solana. But you've still got 92% single client dependency, no native privacy, and a monolithic architecture that can't independently configure privacy and performance. And then you have Ethereum L2s. Uh they're still the largest developer ecosystem in crypto with the EVM, so over 30,000 monthly active developers.

Uh battle-tested L1 settlement, modular, open source. You know, the gap there has always been enterprise tooling, and that gap is closing fast. So, what do institutions actually need? You know, not what we think they need, but what they tell us in the room today. Um you know, we spend a lot of time with custodians, uh centralized securities depositories, who are tokenizing, you know, hundreds of billions in commercial paper, uh correspondent banking networks. You know, what they want is they want privacy by default. You know, when you have um CSDs that run multi-party settlement, participants should not be able to see each other's positions, you know, that's just table stakes.

Uh they want 24/7 atomic finality, and it needs to be cryptographic, not probabilistic. Uh they want open standards, you know, because a lot of these institutions invested, you know, millions in Corda, and then R3 pivoted on them, and they're not going to make that same mistake again. And then in procurement, the number one objection that we that we hear is, you know, will you be around in 5 years? Right? They want longevity. Um they also want developer talent. They want quantum readiness. And they also want interoperability between private ledgers.

So, nobody delivers all of this today, and that's that's the gap. So, here's the thesis. Uh L2s are the only architecture that can deliver all three. You know, privacy, interoperability, and security. Uh private chains sacrifice, you know, trustless interop. Public L1s today sacrifice privacy. Uh L2s can actually resolve that trilemma in a way that's expedient. Um and now I'll show you one how that looks at an architectural architectural level. Okay, so just a quick primer for those that's familiar with ZK or zero knowledge. Um zero-knowledge proof yeah, lets you prove a computation was correct

without revealing uh the underlying data. Um ZK gives you mathematical certainty with no trusted party and no data disclosure at all. Why do institutions care? Because you can prove compliance without exposing positions. You can verify a counterparty executed correctly without seeing their book. You can coordinate across ledgers without sharing state. And regulators can verify correctness without ever touching the raw data. So, that's the primitive that makes this whole architecture possible. How it works. So, on the left you have private execution. So, transactions run inside a permissioned Hyperledger Besu cluster with role-based access control

and policy enforcement. Uh only authorized participants can see the data. Uh the sequencer or block builder uh enforces compliance before a block is even produced. In the middle, you have ZK proving pipeline. So, a prover takes the execution trace and generates a cryptographic proof that everything was executed or computed correctly without revealing any of the underlying data. Uh now you have 128-bit quantum-safe security, and you know, that's not on the road map, that's that's there today live. And on the right-hand side, you have a proof that settles on Ethereum L1, you know, and over $400 billion of in economic security backing it up.

So, what you end up with is the privacy of a permissioned chain, the settlement security of Ethereum, and the verifiability of uh zero-knowledge cryptography. So, that's what Public Chains, Private Guarantees actually means in practice. It's an architecture, not just a sort of slogan. Uh if you compare that to, say, Canton, for example, where settlement relies on their CSP trust model, and Solana, where it's validator consensus, you know, this is backed by mathematics and Ethereum's economic security. So, for a regulated institution, that's a completely different risk profile.

Now, private interoperability. Um I'd argue this is the single highest-value use case in capital markets right now. So, today when two CSDs need to coordinate, say, settling a cross-border securities transfer, uh they use bilateral APIs. Okay, each side keeps their own records. Uh when something goes wrong, and it does, uh you get polling, timeouts, uh manual intervention. Uh the actual post-trade reconciliation alone costs the industry between $5 and $10 billion per year.

Okay? Whereas the proof ba- the proof-backed architecture changes this completely. Uh each participant keeps their own private ledger. And so, when settlement happens, a ZK proof confirms everything was done correctly, and that proof gets shared through a common proving layer. Okay, so both sides get cryptographic evidence instead of a promise. And this is the part I really want to land is that this is not a shared database. Okay, that's what the permission chain sort of vendors promised back in 2018, 2019, but it failed because nobody wants their competitors to kind of see their own see the state.

Um So, what we're talking about here is really just a network of private ledgers connected via proofs. And only the proofs cross boundaries, not the raw data itself. So, you get delivery delivery versus payment can become atomic. Uh both legs complete or neither does. And you get to go from a I trust or API response to a I can mathematically verify your execution. And then what about compliance? So compliance here is layered not bolted on after the fact.

Role based access control sits at the sequencer level so compliance gets checked before a block is even produced. Privacy is configurable at multiple levels. Who holds what token? Who transacted with whom? Where does data live jurisdictionally? All of that can be configurable per asset class. And then on the privacy roadmap, you know, what we have today is a you know, role based access control plus sort of validium but you know, next you can have zero knowledge selective disclosure. So you can prove compliance without revealing transaction data or details.

You know, longer term we can even use tools like fully homomorphic encryption for encrypted you know, computation on chain and then each each stage builds on the last without having to re-platform. And then what about decentralization? So decentralization can also happen in phases. You could start with a single sequencer and you can be in production in you know, under 90 days. You can then move to a multi validator QBFT where the member institutions can actually run their own nodes and then full decentralization can happen over time.

For multi jurisdiction deployments is where you can run dedicated sequences per jurisdiction with native L2 to L2 interoperability between them. Shared proof aggregation as well can make sure you keep costs sub linear as you scale. And on quantum resistance as well the ZK proving I described earlier can be 128 bit quantum safe with an algorithm agility built in so you know, you can evolve as post quantum standards mature. Okay, and then what about the execution client? So this is something that most people don't think about but it matters a lot when you're making a 10-year infrastructure decision.

So most sort of technology stacks today run a forked execution client and ethereum itself has accelerated actually hard forks now every 6 months. So you know, petra fusaka gloucester down with many more to come and every time that happens, you know, every fork has to be manually reintegrated. You know, reintegrate those changes which can take up to 12 even 24 months you know, we've seen so whereas on our technology stack you know, we run the main line besu client. We don't fork it. We build it. We maintain it.

So when ethereum upgrades you know, we're already there. Now on the prover side particularly the risk five piece. This is where it gets really interesting. So our prover is moving to risk five which is a open instruction set. What that means is that you know, we can prove any program compiled to risk five including besu itself as a guest program and that means including ethereum L1 blocks. So we have the same proving infrastructure for layer two execution and L1 state transitions. Why does that even matter? Well, you know, institutional L2s can now natively verify L1 state which means you get trustless bridges. You get native roll up integration. You get full chain

cryptographic verification and no other technology stack can really do this today because no one no one else is sort of running the main line ethereum client. Okay, so the full enterprise the full enterprise L2 stack. So I won't read every line here. You can see the the stack on the screen but the thing I want you to take away is that every component here is built in house by one team. So execution proving identity the gateway. This isn't assembled from a bunch of ecosystem projects. You know, and the gateway as well is worth calling out specifically because it does the ISO

20022 translation which bridges the swift messaging system and on chain settlement. So that's what actually makes this kind of real for banks in production. Deployment as well is flexible. You can have you know, L2. You can have a validium. You can have an L3. It can be you know, managed. It can be self hosted but what's important is you get one team one stack deployed however you need it. Okay, so there's some use cases here and you know, these aren't hypothetical. These are like driving real procurement decisions today. I'll just highlight three. So tokenize securities one European CSD alone is looking to tokenize 300 billion dollars in sorry 300 billion euros in commercial paper.

You have cross CSD settlement which reflects the architecture I just shared and correspondent banking as well. So cross border payments where each bank only sees the transactions that they're actually involved in and all the other stable coins collateral management and custody as well. They're all sort of active conversations that need the architecture I've been describing. I think this table kind of speaks for itself. I think canton is sort of ahead on privacy today. Solana is winning on speed. Ethereum L2s win on pretty much everything else and and the privacy gap is closing fast.

So when it comes to enterprises making 10-year infrastructure decisions, they're not just optimizing for raw throughput. They're optimizing for ecosystem for vendor lock-in for settlement guarantees and on those dimensions you know, L2s win. Okay, so in terms of the upgrade path right like here's the inside that kind of makes this land. You know, a lot of the world's largest financial institutions already run besu today like 80%. So they've already invested in EVM contracts. They've already invested and trained their teams.

They've already built out operational processes around this. So they don't actually need to re-platform. All right, they just need to upgrade. So you can add ZK proving. You can add interoperability. You can add ethereum settlement. You know, you have the same client and the same contracts but now you have new capabilities and so that's not really a migration pitch. It's more of an upgrade pitch and it kind of meets institutions exactly where they are already. Why should everyone in this room care about it? Because of the flywheel, right? Ethereum is trust where and eth value accrual. Every enterprise L2 that settles on ethereum pays gas which increases block space demand which strengthens the economic security and then more value settled basically means

a higher cost to corrupt the network which actually attracts the next institution and it compounds from there. So you know, the numbers here are staggering. Global securities markets are 120 trillion dollars. Our trust where report at consensus models is out in detail and there's a link there if you want to go and check that out. But I genuinely believe that institutional settlement is the most important growth factor for ethereum over the next decade. And the window is closing. So let's be direct about the stakes here. You know, if L2s deliver ethereum becomes a settlement layer for global finance. That's trillions anchored to L1 and here's what gets me excited is that you actually have defi and tradfi end up sharing infrastructure. So imagine a

tokenized treasury from a global custodian uses collateral in an RV pool you know, in the same atomic transaction and that world only really exists if we get this right. So yeah, who's building this? We're obviously building this at consensus with a bunch of other products that we built over the years. It's more than a decade of core infrastructure that we've been building and you know, now we're taking all of that and we're building the enterprise L2 stack you know, which is full stack main line besu quantum safe proving all hosted in open source incredibly neutral way because institutions need to make sure that this is open source and neutral enough for them to audit and trust over the long term.

Okay, I'm just going to skip through because I know we're over time. Okay, yeah. So just to wrap up. Yeah, if you're building in this ecosystem come find me. If you're on the institutional side evaluating infrastructure. I'd love to have that conversation with you. I'll be around for the rest of the conference and my email is on the slide if you want to reach out. Thank you very much. >> [applause] >> Do you have any questions? Well, thank you.