The Elastic Network

Transcript

[Music] [Music] Hello everyone. My name is Alex Kulovski. I am the founder of ZK Sync and today I want to talk to you about the elastic network. uh explain what it is uh what where we are today and share some exciting uh news from the last couple of weeks and and and show the direction where we're going. So what is Zik Sync? At Zikync we're building verifiable blockchain infrastructure for Ethereum. We're extending Ethereum with uh zero knowledge proofs. We're scaling it and we're adding more capabilities.

Uh verifiability is what enables us to build systems that cannot be evil. And this is very critical to our long-standing mission of advancing freedom for everyone in the world through mass adoption of blockchains in the right way such that you don't depend on on humans in the loop and and untrusted parties and um the uh Ziki sync journey started by building a single chain Zikysync era but we since expanded we've quickly realized that um uh the demand for general purpose chains is going to be far lesser than the demand for specialized infrastructure.

There's going to be multi-chain world where every app eventually might become its own app chain. Every enterprise will need its own chain. And thus we open sourced our uh blockchain building framework as Zikync stack and offered it for others to build their own infrastructure and then we expanded it with further things like Zikync SSO or embedded wallet white labelled um front-end solution for for these chains. We added Zync previdium recently the privacy solution. So we we're working on a couple of other things um that can be um explained or described as the all-in-one platform. It's a full vertical integrated blockchain

infrastructure platform with rollup stack fully integrated prover fully integrated uh wallet all the packaging of infrastructure from service providers in like roll up as a service option uh privacy and also very importantly native interal protocol. So I'm going to talk about this. Uh today the ZK sync ecosystem is powering more than 16 chains and more are in the works. Uh the collection of these chains is what we call the elastic network. It's a neverexpanding network of verifiable blockchains natively connected and fully integrated. Um there are some consumer chains, there are some uh interesting

financial chains, social applications like Lance, but the main focus for us today is on enterprises and institutions. Uh because with the recent developments in the US politics, I think it's clear that the um financial institutions penetrating real world businesses is the biggest opportunity we have in crypto in 2025. This is what enables us to move in like make crypto mainstream for real for the first time since the division was born 10 years ago. Um, and we're really excited about this. This this is the um this is not to bypass I our cipher punk roots but bring

them and make institutions, enterprises, companies, governments verifiable, trustless, turn them into systems that move from don't be evil principle to can be evil. Uh we have a number of uh uh already integrated enterprise uh corporate chains that uh add up to over 300 million total addressable user market with some of the largest exchange in the world, some of the largest banks being built on on DK sync. There are more in the works that that I cannot yet share. More things uh are coming here.

Uh and we are also the number one uh sorry number two largest RWA ecosystem in the world today second only to Ethereum with over two billion uh RWST uh on Ziki sync and uh we're first and foremost a technology product company. We're having a very ambitious road map for this year which we shared uh uh at the end of 2024. I'm happy to report that half of the road map have has already been achieved and the other half has made very substantial progress and um I want to focus on a couple of things that we announced just this last

quarter in Q2 2025. Um the first upgrade was the EVM equivalent for Zync chains. This was a huge huge unblocker for the ecosystem. We initially started uh um I think the uh you know we started very early working on zero knowledge proofs the we we exist for for seven years now and we did not anticipate how fast the development of zero knowledge like the the maturity the progress in in zero knowledge proof space is going to be. So we initially assumed that to get to full EVM equivalents it will take several years and we started building specialized infrastructure with a special language then EVM compatible

compiler. Uh we are now at a point where everything can just work completely out of the box and we ship that um uh I think two months ago for era and this is being rolled out for all the other zing chains. uh we could use the developments of our compiler to make a bigger contribution to the entire Ethereum ecosystem and we uh we like transformed our internal compiler for from the ARVM back end into the EVM back end. And uh I'm happy to announce that we have the first ever fully functional end to end working LVM compiler for Ethereum which optimizes gas way better than humans do.

You don't have to do inline assemblies anymore. And uh it produces EVM bite code. It can be repurposed in the future to produce something more sophisticated like risk 5 byte code once Ethereum fully embraces Falix Vision. Uh and it has some really nice things like we got rid of the stack 2D problem. If you're a developer in the space you you know how nasty the problem is. So like please try it out. This is very interesting. Uh we launched the zing gateway which paves the path for interoperability and much much cheaper zero knowledge proofs like lower costs for for aggregation of the proofs in general. And then we have uh um two really interesting things and I'm going to start with air bender. Uh this

is brand new. We we announced this just last week. Airbender is the fastest uh to the best of my knowledge the fastest risk 5 uh ZKVM and prover in the world. At least for a fact it's the fastest open-source prover the the fastest prover that we could grab our hands on and and try out. uh it by far exceeds the performance of the existing alternatives although they they are not even fully open source yet. Uh Airbender is completely open source from um the the CPU prover GPU prover the entire infrastructure and it enables us to prove a single like use a single GPU to

prove Ethereum block end to end in about 35 seconds on an average Ethereum block. So it varies from like 15 to to 1 minute but an average book is about half a minute. Uh and this is a massive breakthrough. It has five big implications which I want to focus on. Number one is this enables us this year 2025 to reach the escape velocity for ZK rollups in terms of costs. We're finally going to beat the optimistic roll-ups end to end on complete costs from from the proving the the maintaining the infrastructure to also data availability because with ZK sync we're using the so-called state div approach to to data

availability. We only publish the outputs of transactions not the inputs. And for most transactions on that that were uh are used on Ethereum, this means massive savings in terms of uh call data. But we still had this baseline proving cost of 1 cent or like slightly less than 1 cent per transaction. We're now going for 0.01 cent for an average transfer. And of course it depends on on on how much gas you consume, but for an average transfer it's it's roughly this this amount. uh so we will be able to completely beat optimistic roll-ups and then this ZK rollups are going to be the clear winner by the end of this year. The second implication is it paves the path towards

home improving Ethereum and of course Zika sync chains. And this is a massive massive uh breakthrough because there are maybe a few hundred large data centers in the world where you can be doing GPU proving but there are hundreds of millions of spare GPUs in households around the world. If we can make proving really really cheap like really lightweight without requiring a lot of coordination, we can distribute these jobs broadly. We're entering the age of zk abundance. The computing power is not limited anymore by what you can rent on on fast AI or GCP or AWS. You can just send it and and we get rid of all the

centralization concerns around proving. So we are we're we can finally be like living the Ethereum vision with full verifiability getting to prove singularity where Ethereum is fully verifiable as sin chain that can be verified on a single smartphone like every single smartphone in the world and proven really really fast cheap uh abundantly on the home machines. The third implication is it offers a lot more freedom for the ZK app developers for the builders of verifiable apps and I believe that eventually every app will get upgraded or almost every app will get upgraded to be a verifiable ZK app

uh because it always has some component that uh requires trust and a lot of these components can be replaced by cryptography. So you can think of things like ZKTLS, Oracle data directly from the source providers that can be verified through uh zero knowledge proofs and then you produce the proof and and of course you have the privacy aspects in the apps and so on and airbender being the fastest open source prover by far by like by a large margin orders of magnitude faster than any nearest open-source fully open source uh alternative released under Apache and an MIT license is a big um unlock for uh for Zeke application developers.

Uh the first thing is privacy goes not just for the uh like privacy is important not not for the only for the end users but for a lot of enterprise and institutional customers we're currently targeting we're working with and I think the Ethereum as a whole should be very very actively going after these days and a big problem they had was that the combination of zero knowledge proofs and privacy was hard because if you if zero knowledge proofs are expensive to generate them uh you either have to maintain your own data center with cluster of GPUs or you have to send your sensitive customer data

into the cloud where you don't know if you can trust them and of course there are some workarounds like tees uh but they're clunky they're they're not quite there and they're not they don't have the same um guarantees as running your proofs on premises and uh sometimes they just like regulatory concerns will not won't allow the this uh these companies to to move um uh and send data to to someone who is not licensed who does not have the same uh infrastructure security etc etc. Uh and this is why we're really really excited about previdium. So previdium is one of the launches we also announced uh in the last couple of weeks. Uh this is a

modification of ZK sync stack specifically for private confidential chains. So you can have a chain that is um fully controlled by you all all aspects of the chain including data availability. So you you have the full control over the data flow. You don't have to disclose it to the world. Uh you can selectively disclose it for example to your customers if you're a bank and you're running your own your own private chain. You can nevertheless have full verifiability guarantees from Ethereum because you're settling on Ethereum and every single transaction is verified through the batched zero knowledge proofs by the underlying settlement layer and uh you also have

interoperability with other private chains in a completely private manner and with uh the public chains. So you can you have access to Ethereum liquidity, you have access to liquidity from other L2s and so on. So this is a massive massive unblock for institutions that enables use cases like uh bond issuance, foreign currency trading, crossber payments, reuse of collateral uh real world assets and so on and so on. Like if we want crypto to to go mainstream, privacy is an absolute must. And what previdium enables is programmable privacy without um custom specialized uh knowledge around uh

special um like domain specific languages for privacy. You can program everything in plain EVM. you can have um um uh you can reuse the existing best practices, the existing libraries, all the tooling that already exists for EVM and have full chain level privacy control over your infrastructure. And the uh the final piece of uh uh the the final implication of air bender is it's a quantum leap for for what we call universal interrupt. makes universal interrop fast and this is a really really interesting aspect so I want to focus on this a bit longer

interrop is the economic rail of crypto all things you know we can start with the basic premise the what we're building is really the internet of value all things that touch value in any form will eventually go on chain from um payments to stocks to any kind of assets to any kind of property titles and so it's going to be a massive trillion multi-t trillion dollar economy by uh 20 2030 but in in the next couple of years cannot really project but if we're building the internet of value the logical consequence is the internet cannot run on a single server and the

internet of value won't run on a single chain there will be a multi-chain world for various reasons for technical reasons but also for regulatory reasons for business reasons different communities just will want to have their own chains and so on and so on. Every company will have its own chain. Every app will eventually probably have its own chain and if we are going to live in the world of hundreds and thousands and more of chains, the interop really comes into forefront. Like can you have the interop in the same way as we have interop on on the internet where every website is connected in one click to to potentially any other website. uh and we can't today with Ethereum the existing solutions are either compromising trust

so you basically like you completely negate verifiability and blockchainness of the solutions with third party custodial bridges which are essentially multisix or they compromise speed and user experience like everyone has tried to move funds with bridges it's complicated you have to think which network where where do I go which bridge I use it it's a mess. Uh and more important most importantly probably it's limiting what you can do. Uh so with with bridges the the trust is the biggest problem with um something like intense which is the the most the closest to what we have to to solving

this problem today. Uh all you can do is send assets. So you can say like I want to swap this asset uh against some other asset and then you still rely on someone to pass the messages that guarantee that the uh something happened on the other chain which you can kind of delegate to liquidity providers. They take the risk for you but then for in the trustless environment all you can do is transfer funds nothing else. So uh our vision for the elastic network the the very reason why elastic network was started in the first place is this idea of universal interrop uh we see it as based on three pillars.

It must be bridge free no trust assumptions no capital parked and and like additional cost of capital like no extra trust assumptions other than the blockchains you're using it must be universal. So basically yeah uh like yeah if I go back bridge free universal and and instant and I I'll talk about them in in in like in detail um the bridge fee means no trans assumptions no capital requirements so like the the connectivity is basically you want a room temperature superconductor for interrop you don't want friction you don't want to be losing a percentage of

value on every single hop which is what's happening today with classic bridges because you have to pay for the cost of capital that you deploy on every single chain that connects the bridges. Um, and you want it uh fully trustless. You want to rely on cryptography and the properties of the blockchains and and your self-verifiability of the systems rather than on someone else's multisc um universal means that it should work for any type of action. any token, not just the most liquid and funible tokens should work for NFTTS should work for the entire long tail of shipcoins or you know some some like

edge fringe tokens that are only used in specific environments to make it a fully seamless experience and it should work for any chain. So it would should work within the ecosystem but it should all also ideally expand beyond a closed ecosystem like an elastic network and uh it should be fast. Um and what airbender enables us to do is to connect every chain in the elastic network through a native interrop protocol with close to 1 second latency. The way it works is every chain in elastic network produces a proof of validity of the

chain and we can produce this proof now in in under one second. If you go to the uh Airbender announcement, there is a fun little game um deployed on the uh on our new test net that is powered by uh Airbender and transactions there confirmed all in all end to end in one second like from the moment you submit it through like inclusion in the block to the block gener like block proof generation to the settlement of this block. All in all, less than 1 second. And of course, those blocks are small, but they're supposed to be small. On Zync, we are the the new upgrade which is coming, I think, next week, is bringing the latency the the uh

the block latency to 200 milliseconds. So, they they will be like closing very fast and the proofs are fully parallelizable. So, if you like we're running this this demo on just a few GPUs. Of course, you can provision as much as many GPUs as you need to to like get to any uh latency that you're targeting. And this enables us to connect all the chains in in a manner where um you deploy once your application on a single chain and then you can use it from anywhere with one second latency without trans assumptions and with full capital efficiency. Uh there and yep that's that's it. Thank you. I'm happy to

answer any questions. [Applause] Okay, no time for questions. Okay, I I can meet you outside. Happy to answer that. Thank you all. Enjoy the lunch.