Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - The Multi-Chain Avalanche - Emin Gün Sirer

Episode Date: May 26, 2025

We couldn’t miss the Avalanche Summit where we sat down with Emin Gün Sirer, co-founder & CEO of Ava Labs, to discuss the evolution of Avalanche’s ecosystem and how HyperSDK ushers in a scalable, interoperable, multi-chain future. Join us for a fascinating discussion on Avalanche’s approach to decentralisation, scalability & interoperability, and learn why real-world applicability relies heavily on bespoke blockchain solutions.Topics covered in this episode:Gün x Epicenter bromanceHow the Avalanche ecosystem & community evolvedThe future vision for AvalancheScaling Avalanche and the HyperSDKInteroperability and the multi-chain futureDecentralising Avalanche L1sUse cases and institutional adoptionOn-chain privacyAvaCloudThe impact of AI in cryptoGün’s focus in the near futureEpisode links:Emin Gun Sirer on XAvalanche on XAva Labs on XSponsors:Gnosis: Gnosis builds decentralized infrastructure for the Ethereum ecosystem, since 2015. This year marks the launch of Gnosis Pay— the world's first Decentralized Payment Network. Get started today at - gnosis.ioChorus One: one of the largest node operators worldwide, trusted by 175,000+ accounts across more than 60 networks, Chorus One combines institutional-grade security with the highest yields at - chorus.oneThis episode is hosted by Brian Fabian Crain.

To scale, you have to have this many chains working as one notion. And that's been proven. And there is none other than Avalanche that has really pushed that vision forward. And now what we need to work on is single chain performance. That's what I'm spending a lot of my technical effort on these days. We need to get a single chain to work super, super fast without compromising any crypto principles, without cutting corners, without doing unsound things. And the hyper-SDK effort is a construction case.

so you can build a specific bespoke VM for your use case. And as a result, we've been talking to a lot of Wall Street firms, and there's a lot of demand for walled gardens. These people have assets that are highly regulated, and they have to have certain requirements met and enforced, and they need bespoke solutions that meet those requirements. If you're looking to stake your crypto with confidence, look no further than course one.

More than 150,000 delegates, including institutions like BitGo, Pintera Capital and Ledger trust Corus won with their assets. They support over 50 blockchains and are leaders in governance or networks like Cosmos, ensuring your stake, is responsibly managed. Thanks to their advanced MEV research, you can also enjoy the highest staking rewards. You can stake directly from your preferred wallet, set up a white label note, restake your assets on eigenayer or symbiotic or use their SDK for multi-chain staking in your app. Learn more at Chorus.1 and start staking today.

This episode is proudly brought to by NOSIS, a collective dedicated to advancing a decentralized future. NOSIS leads innovation with circles, NOSIS pay, and Metri, reshaping, open banking, and money. With Hashi and NOSIS VPN, they're building a more resilient privacy-focused internet. If you're looking for an L1 to launch your project, Nosis chain offers the same development environment as Ethereum with lower transaction fees. It's supported by over 200,000 validators making NOSIS chain a reliable and credibly neutral foundation for your applications. NOSISDAO drives NOSIS governance where every voice matters. Join the NOSIS community in the NOSISDAO forum today.

Deploy on the EVM-compatible NOSIS chain or secure the network with just one GNO and affordable hardware. Start your decentralization journey today at NOSIS. Cool. Well, it's so great to see so many people here. And what a beautiful space this is. And yeah, so my name is Brian Crane. I've been in Crypto for a long time. Actually, in 2013, I started a podcast called Epp Center, which we are still running. And at this point, it's the oldest crypto podcast. And one of our most frequent guests, pretty much almost at the top of the ranking, is Emin Gonsir, who we've had on so many times. actually from way before Amalange even when... Way before it goes back a long ways, all the way back to selfish mining days. Selfish mining, Bitcoin NG. So he was at the time, he was professor at Cornell.

And he was doing some of the most interesting papers really on an academic research on Bitcoin and cryptocurrencies. Yeah, when we were getting attacked by the Bitcoin Maxis for all the scientific work, the epicenter podcast was the one that actually gave us a platform to get the... story out and they were the most scientifically driven podcast out there. So it was always a welcome thing to go to. It's wonderful to have him here with us at the summit. So, you know, there's a, there's a lot of bromance going on between the two of us. Always has been. So also the avalanche paper now came out seven years ago. I think Eva Labs, you started seven years ago as well. It's six, yeah. Six years ago. Of course, now, you know, big conference.

here, so many people, like a whole ecosystem has emerged. Like, when you look at the Amalanche ecosystem today, what are you most proud of? So, yeah, that's a great question. I'm really, really proud of the community. That's, I think, what we managed to build. Of all the things we did, you know, we did a bunch of technical decisions, a bunch of new technical developments. We brought a lot of, a lot of new science into the space, and we ended up delivering that which other people promised, right? You know, when we started, started, Ethereum was proof of work. They didn't know how to do the transition. We showed them how that it was doable and that you could build really fast finality systems. You know, by the time they pick it up, you know, they get two and a half minute finality and we have sub-second finality. But, you know, we showed, we paved the way. I'm really proud of all of that. Sure. But really it's the community that we managed to assemble that brought us all of these new avalanche L-1s, that brought us all of these DFI protocols, all of these games. all of these exciting developments that are happening on top of us.

I think that's really the thing that I'm most proud of. And that's really the thing that I spend most of my time worrying about, right? You can have the best technology, but it has to be coupled with the people that really push it. And in getting the word out, building that community has been, I think, both the greatest challenge and probably my proudest accomplishment. Cool. And now looking into the further future,

if you think like five, 10 years ahead, what do you want the avalanche ecosystem to look like? So I think it's pretty clear where we're headed, right? We showed the world that to scale, you have to have this many chains working as one notion. And that's been proven. And there is none other than avalanche that has really pushed that vision forward. So we showed everyone how to do it. Interoperability should be built in from day one. And you can't fracture.

you've got to have an architecture that doesn't fracture you at the social level. And I think it's been interesting watching other chains make those mistakes. So where are we going with this in five to ten years? Well, we showed everyone how to do the multi-chain thing. And now what we need to work on is single chain performance. That's what I'm spending a lot of my technical effort on these days. We need to get a single chain to work super, super fast without compromising any crypto principles, without cutting corners, without doing unsound things like loading your state from Google Cloud.

Can't be done. Like cutting out verifiability of the chain. You can't do that. There are a whole bunch of things that other people have done to get speed. And I believe that we can build things that are faster than what they've done without actually doing those compromises. So that's, I think, where we're headed. And at the sort of the scale above us, at the level above us, at the political level, there are a lot of exciting changes happening.

Boomers have now picked up on coins, right? So we're seeing this resurgence of interest in Bitcoin. Fantastic to see. We're seeing politicians, you know, get excited about meme coins. Wonderful. But then the next level up from that is going to be programmable money. And they haven't picked up on it, right? Defi, like Trump doesn't know what defy is, right?

That's he hasn't gotten there yet. And it takes people like, what, six months to a year, maybe two years, if they're savvy. So we're going to get there. And when those changes happen at the societal level, that's going to unleash another spurt of growth for the entire ecosystem. Companies ought to be building their business flow on-chain. Defi and financial rails should be powering everything of value.

So that hasn't happened yet, and that's coming. That's what I'm super excited about. So you talked about the scaling of the L-1s, I know there's a lot of work that has gone into hyper-SDK, right? So having a special SDK for building your purpose, special purpose chain. Of course, today, most activity on Avalanche, and I think in many ecosystems happens on EVM, what do you think is the role of these two in the future? Do you think mostly it will be hyper-S-DK or it will be EVM or EVM on hyper-S-DK?

How do you see that play out? So for those of you who don't know what hyper-S-DK is, It's really a construction kit for building your own virtual machine. So those of you who've done, how many people here have done solidity programming? That's a fair number. So if you did solidity, how many people have done solidity on EVM? It's probably the same number. And how many people have done solidity not on an EVM?

Nobody, almost, I think. Very few. So as you can see, the default path that gets people programming right, smart contracts is one based on solidity for the EVM. That's why when we started, we decided, okay, we're going to adopt this. One of the hardest paths to change is the social path by which people get into the space, and that path directs them to the EVM. So we will inherit the EVM with all its flaws. It's not how I would have designed a fast virtual machine. They made a lot of mistakes there. But we'll take it and we'll show the world how to build it faster, and that's

exactly what we're doing. But for those people who have highly demanding needs, we wanted to provide a way for them to build their own virtual machines. And the hyper-SDK effort is a construction kit. So you can build a specific bespoke VM for your use case. So if you have a very demanding decks, if you have a very demanding game, for example, you can build your own virtual machine with its own op-codes, et cetera. This is, it's hard work. This is beyond the scope of most normal programmers, and it should be. It's a big undertaking. So what's going to happen moving forward?

Well, I think exactly what has been happening. We're going to continue to support the hyper SDK and the construction kit, but we don't expect most of you to use it. And as you can see, most people are not using that. You don't need it. And the learnings that we pick up from the hyper SDK effort, from the construction kit, we want to take and apply

to building a fast EVM. So the EVM has, for all its flaws, has a lot of popularity and it's possible to make it go a lot faster than what's been built so far. And we want to use the techniques that we have built into the hyper-SDK effort and adopted for the EVM. You mentioned the issue of intraoperability, right?

So it has been one of the big challenges, right, of application-specific chains is that you have some fragmentation of liquidity of the user base, can be user experience issues. And so today still, right, we have this big contrast of visions where on the one hand you have the vision

of like, you know, many L-1s that, you know, probably have some kind of interoperability standard versus, you know, single high performance maximum throughput chain, you know, you see like Salon, Monad, others pursue. Do you, Like, how do you see this play out in the future and do you think we'll have both?

Or do you think we will have also applications where you don't even notice it's on different chains? And the user explains is so extracted? So let's see. So this is a great question, right? What's going to be the multi-chain future of the world? And so let's see. In the short term, we're going to continue to see what we've been seeing, which is these new chains are going to come up. They will have specialized use cases. Solana has attracted the meme coin crowd, for example,

and that's a fine place for them to be. It's okay if you want to go to the meme coin casino, then everybody pools themselves at some location that's chosen. It doesn't matter how it's chosen, but you want to be with other like-minded people you go there. That's fine. But in the long term, the network effects are so strong that we will see. Okay, let me actually go back. back in the 70s and 80s, when networking equipment was being pushed, there were all sorts of contenders. There was data kit. There was, you know, all sorts of proprietary protocols for connecting computers. And then came the Ethernet. And now everyone at home has Ethernet cables, right? You never hear of anything other than Ethernet.

It's just over time, it's going to be a one system takes all. And how did Ethernet succeed? It was cheap, easy, fast, universal. reversal and it was interoperable. So that's what we're building. It should be very, very, very easy to build your own chain on Avalanche L1, and it is. It should be very easy to have it interoperate, and it is. And it should be seamless to connect, and of course it is. So in the long term, if the network effects just create a funnel towards you, you're going to be the dominant winning player. And that's the big game that we're playing for. So in the short term, yeah, of course it's stuff will come up,

Monad will come up, I don't know, all sorts of new things will come up with a fancy thing, and then they'll fizzle out. A lot of them fizzle out. But the big winners here are going to be those that do the Ethernet play properly. And that's exactly what we're going after. For Avalanche L1s, I'm curious. So today, a lot of them launch in a fairly centralized way, right? And then I think there's the idea that with time, they decentralized more.

like, what do you think is the right degree of decentralization and what time horizon for the L-1s? That's a great question, Brian. So let's see. And we can ask the audience, right? So what should be the right level of decentralization? Like, think about that for a second. And if we sort of, you know, kind of like look into, you know, I don't know, like look at like whatever application you might have in mind, the level of decentralization you demand will vary. there is no way that there can be a single answer to this question.

I don't even know the units of decentralization. Let's call it like one Nakamoto or whatever, right? So you want to achieve some amount of decentralization. You certainly don't want anybody meddling in a system in a manner you don't understand or expect. And you need that. But what does that mean? It's different from a game. It's very different for a defy application, very different for a governmental application.

It's just, it's a continuum. So, once you realize that, then it should be obvious that a single solution cannot possibly satisfy everybody. Like we have, I don't know how many people in this room, but undoubtedly the amount of decentralization you demand cannot all be the same. So if that's the case, then how can you expect a single chain to possibly emerge as the ultimate solution? How can Solana, for example, which is a single chain solution?

How can Ethereum, which is a single chain solution, possibly meet all of these conflicting demands at the same time? It can't be possible. The only winning solution here, the only way to win this game to satisfy all the different needs is to have a multi-chain solution where different chains have a different level of decentralization depending on the use case. A game, for example, starting out might be very very very much. very, very centralized, and then it opens up, and because more decentralized, you know,

governmental applications can be totally centralized and provide only auditability. And then there's a whole slew of solutions in between. So that's how I feel about this. We have to be able to provide to the users the ability to select the level of decentralization that they need for their application. And that's one of the key tenets that helped us when we were designing Avalanche, five years ago, that we need to be able to meet these deferring needs for different applications. So let's talk a little bit about the use cases. At the moment, institutional adoption is a big

thing that's happening. I think Avalanche has always been very early and very deep on the institutional focus. I think Aval Labs is certainly one of the teams in the industry with the best you know, BD institutional focus? What are the use cases that you feel are the most promising and where we're going to see the most traction in next one or maybe three years? So, yeah, that's exactly right. We ended up, I think I mentioned this yesterday. We ended up moving to the most contested, the most regulated real estate in the world, right? We moved to Manhattan, well, Manhattan slash Brooklyn. We moved to New York City. And that's because we wanted to be close to institutions that manage large asset bases.

And as a result, we've been talking to a lot of Wall Street firms, and there's a lot of demand for walled gardens. These people have assets that are highly regulated, and they have to have certain requirements met and enforced, and they need bespoke solutions that meet those requirements. And one of the very common things are certain specialized markets for groups of companies that are cooperating towards some end.

So we're building a couple of these for Wall Street. I think we have announced some of them with J.P. Morgan, with Tiro Price, etc. Those are all been made public. But there's a lot more in the pipeline that I'm really, really excited about. And expect now that there's more regulatory clarity,

expect all of these consortia to gain speed again. Everything came to a standstill at some point. I would say maybe around two years ago. So there was a lot of interest, and then after FTX, everybody took a step back, and after Operation Choke Point, they really slowed down. But now interest is back up again, and I think there's going to be a lot of exciting institutional solutions.

And how many people here have actually looked at the softwares? How many people here have worked in TradFi? That's quite a lot. And so you've looked at the software stack that people are shaking their heads and yeah, you know how messy it is. The amount of effort that goes into reconciliation, the amount of effort that goes into making sure

that my database and your database are coherent with respect to each other, that the trade that I think happened at this time, you also think happened at that time at the same price. That effort is so immense and those layers and layers of software are so complicated and the error case is so, so bizarre that the moment you see the blockchain solution,

he said this is the next level. Like this has to be the new rail. So those are coming. And so those bi-temporal databases, like these databases that have multiple sense of time, like the time I think it's happened, and the time you think it's happened, and all of that complexity needs to go out the door.

Everybody realizes this. And those crazy layers of software just to maintain some illusion of coherence between two different parties, it has to go away. People understand the value of blockchains and auditability, of course, as well. So it's coming, and institutions are excited again.

Regulatory clarity will clear the path. The Genius Act is making its way through Congress. So I'm thrilled about what's to come. The next four years is going to be one of regulatory clarity and opening up on the social side. And I think the era of constantly waiting for some shiny crippling, You know, cryptography object, is over.

Like remember how VDFs, verifiable delay functions were supposed to save the day? Remember how verifiable random functions we're going to allow Ethereum to scale. And then accumulators were going to come in and that didn't work. And in other words, ZK EVMs were supposed to save the day. Those are gone. The days of selfies with Vitalik, much as I love him and we're very close friends, those are gone. and I think he's relieved as well, right? So we never played that game,

but a lot of other people did. And, you know, and I have to admit, what have those days been replaced with? Now it's time for selfies with Trump, right? So boomers are getting into the game and they're slowly figuring their way out. But in about a year or so, it's going to be the time of actual use cases.

And I cannot wait, because that's where we're strongest and our ties, our decision to go into where the assets are in New York has been an amazing decision. And I cannot wait the results of those efforts. So for these efforts, right, where you bring traditional assets on chain, I think everyone probably who has, you know, use crypto for a while can feel the advantages of it.

But one of the big challenges, I imagine, is around privacy, right? Because today still, right, you can, there's actually a lot of privacy in the traditional financial system or a decent amount. on-chain is a lot harder. What do you think the privacy requirements are going to be to bring finance fully on-chain? So so far, until recently, it was very binary, right? Either you worked on a public chain and everything was visible to everyone,

or you used something like Z-cash solutions where everything was completely opaque and often there were bugs and so forth and you couldn't even audit how many Z-cash were outstanding, etc. So there was no middle ground, and that's not a good situation. But recently, we released something called encrypted ERCs. So this came out of our work with various different institutions again, where people wanted to build consortia.

But within the consortia, they want to cooperate, but they don't want to reveal how often company A does business with company B. But they wanted to be auditable to a certain set of parties. or in Wyoming, for example, again, they want to be able to use stable coins for a lot of payments. They want the regulator to have full transparency into what's happening, but they don't want the participants to see exactly how much money everybody else has, exactly how much money they're sending to each other. Insanely reasonable requirements, and there was no way to meet them.

So to respond to that need, we developed something called encrypted ERCs. And there, there's a level of privacy that's granted, to all of the players. There's an auditable chain, and there's an ordered timeline, a way to go back to the beginning of time and verify all of the results. And yet there's a way to mask amounts

and senders and recipients from each other, if necessary, and a way to expose everything to regulators where necessary. It's all adjustable. So that solution is coming, and it's one of the big differentiators, and it's one of the main reasons why we've been fairly successful meeting the needs of these highly demanding financial use cases where privacy is paramount. So that will be on the C-chain and available for any EVML1? Yeah, exactly. I think, I'm not sure if it's on the C-chain today, but all of the EERC code has been made public.

It's under the Avalanche ecosystem license. So you could just take it and launch it on any chain you like. It could be on your Avalanche L1. It could be on the C chain. And then take advantage of all of those privacy features that I just talked about. One thing I'm also curious about. So there is Ava Cloud. What's the vision for Ava Cloud and the role Ava Cloud is going to play in the future?

Yeah. So Ava Cloud is a – so it's important to understand the difference between Avalanche and Ava Labs, right? Ava Labs is the company that provides the technology for Avalanche. We wrote a bunch of the code. as well as, of course, there are a bunch of faces here that I know have contributed code to Avalanche. So Avalanche is a community project where Ava Labs plays a role. But Ava Labs has a separate arm inside that helps people with demanding use cases build their own L-1s and launch their own chains. And so that's OVA Cloud effort.

And Ava Cloud allows for us, for example, it's a revenue-generating for-profit effort. and it's a set of people that go and talk with someone who wants to launch a chain and they understand what they have a process by which they figure out what the needs are and then very, very quickly they can build a custom solution that meets those needs. And if you have something in mind where you know you want to digitize some asset, but you have unique requirements, you know, Alva Cloud can typically very, very short order give you a chain, give you a wallet, give you an explorer, give you everything you might need to have your own L1 according to your own rules. So it's been fascinating so far. I think they have

launched 50 plus chains on Ava Cloud, and they have a backlog of 150, if I'm not mistaken. So there's a lot of effort that's gone into helping other people launch their chain solutions. And for example, like the game OTG off the grid, that's taking advantage of Ava Cloud. There are many others that are Maple Stories, another one that's powered by Ava Cloud. These high demanding applications where somebody says, hey, I know how to do my job, help me with the blockchain side, that's what the Ava Cloud effort has been doing, and it's part of the revenue generation for Aval Labs. So let's talk about the favorite topic of everyone at this point, which is AI, right?

So, of course, AI has really sucked up a lot of the attention in the entire tech space for a while. in crypto, everyone is interested in how does this conversion happens between AI and crypto and, you know, how can different chains, different applications play a role in that. So I'm curious, what do you think is going to be the impact, you know, of AI on crypto and maybe the role crypto can play in the AI revolution? So the very first way in which people took AI and connected it to a blockchain was through these AI agents. And AI agents are big on Avalanche. We love them. It's wonderful to have an agent attached to an address on the blockchain, and then you can interact with that agent. That's awesome.

But it's very centralized. You're dependent on an external off-chain entity to run that AI agent. And so people are going to push that effort. It's exciting. It's interesting. I wonder what will come of it, and we're watching that with great interest. But about two years ago, I had this idea that I actually discussed at the summit in Barcelona two years ago, which was coin-operated agents. So my belief is it's so much better if you're not dependent on an off-chain entity. It's so much better if it's decentralized so you can count on some AI intelligence up in the sky that will be maintained even if there isn't a singular counterparty that's maintaining it.

So, the vision for coin-operated agents is that we will build chains where every validator has its own LLM in it, and we will be able to interact with those chains using natural language. So I think, in the long term, I would hope that we will ditch virtual machines. There shouldn't be an EVM. You shouldn't have to learn solidity. You should just write in your own language. You write in English, I want to give five bucks to Brian or what have you. And maybe you express something far more complicated, right? I want to give so much money to Brian, assuming he can do da-da-da-da, and if not, I want my money back.

So, or maybe something far more complicated. I want to define a lending service where Stani. I'm not sure if Stani is here, but imagine writing the requirements for Avey, right? I want to define a lending service where there are two tokens and one is convertible to the other. It's a bit of a complicated description, right? It's a spec. But it implements a lending service. And you should be able to do it in English, in German, in Tagalog, in Chinese, whatever your language might be.

And other people should be able to interact with it using their natural native language. And that's, I think, where a lot of exciting things lie. So to that end, we've been building a lot of prototypes. Now, many of you are probably looking at me thinking, this is probably a bad idea. I agree with you. Okay, it's a bad idea. So much money can be lost if you say the wrong thing. to the AI, right? And also, are you like not empowering AI to have its own wallet now and it can do

autonomous crazy things with it? That too, right? I mean, there are a whole bunch of issues that this uncovers, but it's so exciting. The end result is a world where you don't have to be a programmer. A world where anyone who can speak a language and can enunciate what they have on their mind can get work done on a chain. That end world is so amazing. because it opens up chains to billions of people who just speak a language. That it's worth exploring and it's worth taking the hits and it's worth seeing what we can do with this potentially interesting and technology with a bunch of potential downfalls.

So I'm really excited about exploring this. We've done a bunch of prototypes at Aval Labs. I'm doing another project this summer. It is clearly not time yet. The AI agents are not good enough to launch this stuff. But we're getting there, and I'm really, really excited about what's going to happen as a result. Maybe let's do one final question. When you think of the next year, for you, what's your biggest focus, priority, the problem that keeps you most obsessed?

Getting sleep. So, no, I don't know what. So there's so many, right? So single, the main thing that I'm working on these days is single chain speed. So I think I talked about Firewood, the database effort, that we built a new database specifically for blockchain applications, and that makes us quite different from everybody else who's using the same stock database.

We have these changes coming to reduce the data footprint of chains so that we don't have to spend gigabytes and gigabytes just collecting data that nobody ever uses again. And then there's a lot of effort to build async, execution of overlapping consensus with execution. That's going to give us another couple of orders of magnitude in terms of speed when you have multiple parallel chains going on the machine. So it's very technical. It's at the very lowest level. And I'm super excited about those. That's been my focus. And I expect to have a lot of exciting code releases in the next six to nine months

to translate all of this into practice. Cool. Thank you so much, Gurd. It was really, a pleasure to once again speak, and I'm sure there will be many more conversations in the future, and I'm super excited about what's ahead for the Avalanche ecosystem. It feels like things are coming into a place, and the foundation's there, great community. So I think there's a lot of exciting things ahead for Avalanche. Thank you so much, Brian. Always a pleasure to chat. Thank you all more.