Scaling Ethereum To The Next Level with zkEVM

Hey Bankless Nation, I'm very excited about the episode today. David is out, and this episode gets technical at times, so I have ETH researcher Justin Drake, whom I'm sure many of you know, he's co hosting with me today. Some context as we get into this episode. So we did a previous episode a few weeks ago called The Sci Fi Roadmap to Ethereum. And in that episode, Justin Drake described the end game for Ethereum. And he said this we snarkify everything.

In today's episode, we explore exactly what that means and how we do it. How do we snarkify everything? Our guest today is Brian Redford. He is the co-founder of what might be the world's first type 1 ZK EVM. And if you don't know what that means, that's fine. Neither did I, as we're getting into this episode. And it turns out that building a type 1 ZK EVM is an important part of delivering what Justin Drake called an enshrined roll-up inside of Ethereum. More on that in just a minute. But before we get in, just want to mention something quick from our friends and sponsors over at So why are we having this conversation? And why now? The compute era scaled with Moore's Law, but the blockchain era scales with something differently. It scales with cryptography, specifically cryptographic breakthroughs like ZK Snarks and all of this ZK Snark stuff, the snarkifying of the EVM, it's all happening a lot faster than any of us previously thought. It's happening so fast that a project called Risk Zero just came on our radar last week and they've already produced a working version of the world's first ZK snarkified type 1 EVM. What kinds of things could this unlock in the future? Why is this important? Well, what if we could convert an optimistic rollup to a ZK rollup? What if we could upgrade Ethereum's layer one from a single-threaded EVM model to a multi-threaded EVMs so that compute was virtually limitless and free? What if ETH validators themselves had the ability to run from something as small as a smartwatch? All of these are possible unlocks with this technology. This is crazy cool, deep stuff. Down the crypto rabbit hole, we go, and it gets technical at times, but it's absolutely worth holding on for the ride. This is crazy cool stuff, and we're going deep down the rabbit hole today. And this gets technical at times, but I think it's absolutely worth it to hold on for the ride. Because this is key to understanding how blockchains actually work and how they scale. And in so understanding, I think this type of thing can help you avoid bad investments and dead ends. And there are a lot of those out in the space as well. We're going to get right to the episode with Brian and Justin, but before we do, I want to thank the sponsors that made this possible, including our number one recommended crypto exchange for 2023. Kraken. Go check him out. Bankless Nation, I am super excited to introduce you to Brian Redford. He's the co founder of a ZKEVM project that we're going to find out a bit more about on today's episode called Risk Zero. Brian, welcome to Bankless.

Thanks. Glad to be here.

Also excited to be joined yet again on Bankless by Justin Drake. He's an Ethereum researcher and repeat bankless guest. Justin, how are you doing?

I'm doing great. Thanks for having me again. But I guess this time maybe as a host asking some technical questions.

Yeah, how's it feel? The tables have turned. So, Justin, uh, I'm gonna tap you in as my co-host today. So, David is out, and we're gonna talk about this uh the next generation ZK EVMs. So I think we're talking to Brian about the world's first maybe ZK EVM. That's a type one ZK EVM, and I'm not even sure the words that I'm saying or what they mean. So we'll we'll absolutely need to define that. But um, you know, David's out right now, so Justin, you're gonna you're gonna tap in and help me with this. I feel like this is a continuation though of a conversation that you had with him on ETHCC. And I think bankless listeners may have listened to an episode entitled uh Ethereum's sci-fi roadmap or the sci-fi roadmap to Ethereum, in which there was this really interesting part, and I love that episode, by the way, where you were describing this uh the ability of us in the future of Ethereum to snarkify the EVM on kind of mainnet, on the base layer. And that sounded really interesting to me, and I think that ties into the conversation that that we're about to have. So tee this up for us, if you will, Justin, as a continuation on that conversation. I know this we're talking about sci fi Ethereum future stuff, but uh, what does it mean to snarkify the EVM? And how does that tie into the conversation we're about to have with Brian today?

Right. So big picture, we're actually going to Snarkify all of Ethereum. And there's two big components that need to be Snarkified. One is the EVM, which is this virtual machine which processes Ethereum transactions. And then the other part is the beacon chain. Now, once we've snarkified these two things, we'll be in a position where compute won't ever be a bottleneck for Ethereum.

So it means that, for example, as a validator,

you won't have to really have B fee CPUs. So you'll be able to be a validator on your smartwatch.

If you're building bridges between L1s,

you'll be able to have another L1

verify the state of Ethereum

without having to redo all the computations themselves. It also has implications for like clients, for what we call enshrine rollups, which are super high security rollups.

And

when the words type 1

EVM come to mind, type 1 ZKVM, it really kind of gets me excited as a researcher because it was a kind of a piece of sci-fi engineering that was thought to be five to ten years into the future.

But it looks like, well, there's several teams working on them.

There's, for example, uh Taycho, and there's Risk Zero now.

And

uh it looks like the engineering will just be ready so much sooner. So we're starting to see the light at the end of the tunnel,

and that has some big implications in terms of applications for Ethereum, both at layer two, but also what I'm excited about, which is the layer one.

So what you just described here, Justin, is uh is kind of the holy grail. And I still want to spend some more time with you right here, just fleshing this out and making sure that that we understand this going to the episode because it sets the context for the rest of the conversation with with Brian and Risk Zero here. So the light at the end of the end of the tunnel, or what I just referred to as as the holy grail, snarkifying all of Ethereum. What this means is we use get to use the spooky math, you know, the crypto magic ZK math that you've you described so eloquently many times on the Bankless episode. And um what we get, the prize that's held out to us, is the ability, first of all, you said, to validate, be a validator on something as small as like a smartphone or an Apple Watch. So, okay, is is that really what we're talking about? So, of course, one of the the end goals and the end goal, the entire purpose of Ethereum is uh

For it to maintain for it to remain decentralized. And that mean means ideally, anybody with a basic consumer level uh computer

can validate transactions on the Ethereum um mainnet. And right now the the requirements for doing that are somewhat higher than just a smartphone or a smartwatch, but this will decrease the requirements, the hardware profile, in order to validate transactions on the Ethereum uh mainnet and also to stake. Is that what you're saying?

That's correct. So anytime you have a human that ends up wanting to interact with Ethereum, they need to interface through a full node.

And

There's some complications here because running a full node is not something that you can necessarily easily do on your phone.

And it's not something that the average individual individual will want to do. And it's something that every single validator has to do if they want to become a validator.

So there's this barrier to entry. And more often than not, today, for the vast majority of users, we end up patching this technical barrier to entry with some level of centralization.

So, for example, if you're using MetaMask,

you're going to be connecting to Infura nodes that are kind of running the full nodes on your behalf. And so there's some amount of trust that you as a user are placing into MetaMask.

So once we have a type one ZKVM, once we've Snarkify DVM and all of Ethereum, we're going to be in a position where the user will be able to interact with Ethereum with much, much, much less hardware.

Like

A phone or a smartwatch will be able to get best in class access with best in class security, best in class latency,

all with very, very little hardware.

Does this imply anything for a bandwidth as well? Will this decrease the bandwidth requirements or will bandwidth kind of become the constraint here now?

Right, great question. So,

consensus kind of solves

two problems. One of them is execution, and the other one is data availability. SNARKS is kind of this magic technology that removes computation as a bottleneck within the context of consensus.

And it turns out we have another magic technology for data availability called data availability sampling.

Now, neither of these are really in production right now.

But once we have both in production, you won't have to pay the cost of computation and you will have to pay a very, very minimal cost from the bandwidth perspective. So you won't even need to download the Ethereum blocks. All you'll have to do is make these small queries for chunks of blocks. And that's going to be enough to guarantee that you're on the canonical Ethereum chain.

So data availability uh sampling as well. That's that's a core upgrade of what I think um people are calling dank sharding as well. Not proto dank sharding, as I recall, but but dank sharding, which which uh could occur later in the future.

Exactly right. So we're looking at technologies, you know, two, three, maybe four or five years into the future, which in a way will transform

accessibility of Ethereum for users. And in the endgame,

accessing Ethereum will be just as easy as accessing any other website. And you'll have guarantees, just like on a website today, you have this cap lock.

Um, sorry, this uh this lock uh NHTPS, you'll have a similar lock saying you're really connected to the real Ethereum and you'll have to have done almost no work and you'll have to download it, download almost no data.

And ZK trust nobody also, unlike the current kind of lock where you actually have to trust the sort of

signature verifiers

for those signature issues.

ZK technology is the thing that that makes this this all possible. What's very interesting is I I know there are people who talk about um kind of like compute scaling in blockchains via Moore's law, and that's true. But where we really get the like kind of the massive uh scalability is more with like cryptography breakthroughs. That that is something I've I've learned as part of like exploring the roadmaps and being in this industry um for uh for for many years now is is these cryptographic breakthroughs are the key sort of step function breakthroughs that allow us to actually scale this technology. One one other quick question for you, um Justin, before we get to Brian to kind of describe what what he's actually working on, what we're doing here. What does this imply for maybe uh scalability on Ethereum mainnet? So we talked about lowering the compute requirements to be a validator, which is fantastic. That is a you know further decentralization unlock. How about transactions per second on the Ethereum layer one mainnet? Does this have any impact on that as well?

Fantastic question. So once we've snarkified the EVM,

we'll be in a position where we can greatly increase the gas limit

and potentially even remove the gas limit

for computation specifically.

And the reason is that the gas limit is an anti denial of service vector, whereby

when you receive a block.

You want to be able to fully verify that the block is valid on the order of one second. And so if there's too many transactions in that block, then it's going to take more than one second to validate.

But the magical thing about SNARKs is that the verification time of a SNORK is on the order of one millisecond.

And so you can take a block that's kind of arbitrarily large with arbitrarily many transactions and arbitrarily much transaction execution