On The Brink with Castle Island - Dan Burke (Nirvana Labs) on Purpose Built Infrastructure for Web3 (EP.558)

Episode Date: September 4, 2024

Dan Burke, CEO of Nirvana Labs, joins the show. In the episode we discuss: Dan's background and journey into the crypto space The role of nodes in a blockchain network and the complexities of running one independently Challenges with existing node infrastructure providers as companies in the space scale Nirvana's product offerings including managed nodes and web3 cloud and the benefits of leveraging bare metal servers Dan's long term view on how the node operation landscape evolves

Welcome to On the Brick. This is Sean Judge. Today, I'm joined by Dan Burke, CEO of Nirvana Labs. We recently announced that we co-led a $4 million seed round into Nirvana alongside RW3, Hash 3, and others. Nirvana has developed cloud infrastructure, purpose built for the crypto industry. Here's my conversation with Dan Burke. Matt Walsh and Nick Carter are partners at Castle Island Ventures. All of these expressed by them or the guests on this podcast are solely their opinions and do not reflect the opinions. of Castle Island Ventures. Guest and host may maintain positions in the assets discussed in this podcast. You should not treat any opinion expressed by anyone on this podcast as a specific inducement to make a particular investment or follow a particular strategy, but only as

an expression of their personal opinion. This podcast is for informational purposes only. Brought down by bad mortgage investments, Lehman, which has 25,000 employees, will be liquidated. The federal government loans American International Group, AIG, $85 billion. This is a different kind of market, and the Fed is asleep. The federal government is stepping it to stabilize Fannie Mae. and Freddie Mac, the two mortgage giants that have been threatened by the housing crisis.

The Bank of England has pumped 75 billion pounds more into Britain's ailing economy with a new round of Concentive Easy. You print a couple trillion dollars and all of a sudden people start to worry. So out of this worry, we have something called the Bitcoin. Bitcoin. Dan, welcome to On the Brink. Thanks for joining me today. Thanks for having me.

Glad to be here. Well, before we dive into Nirvana, it would be great to share with the listeners a bit about your personal background, how you made your way into crypto, and what led you to eventually getting involved with Nirvana? I am born and raised in Cincinnati, Ohio. I've spent the last 10 years in Singapore. What originally brought me out here was I used to be a physical commodities trader. I think there's a lot of overlap with what commodities and crypto do. I was originally drawn to crypto from Union Square Ventures blog post. Bought crypto, held in Coinbase. I'm sure a lot of listeners potentially remember this. In 2017, Coinbase, they had said that they were not supporting

the Bitcoin Cash fork. You kind of had to go out and learn how to withdraw the crypto and learn what a private key was and learn how to sign messages on different blockchains. And it was then that I was like, wow, this is quite powerful. I want to get involved in this. And yeah, I was sitting in Singapore. So all I did was looked at U.S. companies, which were expanding internationally. There weren't a lot at the time and found Bitco had just done their series B applied, had a bunch of chats with them. I was one of their first hires outside of the U.S. I joined with the guy around the same time that set up their UK office. I set up their Singapore office and really just got kind of lucky that Singapore has become one of the leading hubs for crypto. So I was at Bitco for four and a half

years. Started there when there was three, four assets that we supported in hot wallets. And then we did the whole qualified custody thing, prime brokerage, on-site custody. There was a lot that happened in the four and a half years. Took that experience and parlayed it to another startup called Preeta, which was doing a decentralized MPC, which eventually led me to get linked up with Nirvana. And so that's where I'm at now. Awesome. Before diving into Nirvana, if we just step back for newer listeners, in your words, can you

share the role of nodes in a blockchain network and why? companies and protocols need to ingest and understand this data? At a high level, I think there's two important nodes that listeners need to know about, and that's the RPC blockchain node and the validator node. And the easiest way to think about a validator node is they create and enforce all of the new blockchain data. And then RPCs take that data, they ingest it, and then they store it. They store those data in servers around the world for Web3 applications to refer to.

The best way to understand why companies need to ingest and understand this data is probably through real-world examples. So if you're using a crypto wallet, those crypto wallets need to know what your client balances are. And then once you submit a transaction through that wallet, they need to perform and submit the transactions to the chain. You can only do this via an RPC node. And once the transaction is submitted, it goes over to a validator nodes to actually create that data and put it into the latest block. Separately, if you're a data analytics company, maybe you use websites which match key opinion leaders, Twitter profiles with their public addresses. Those platforms are reliant on the latest actionable data from those addresses. The only way you can get that data is from reading an RPC node.

Prediction markets are pretty hot at the moment with all the political betting going on. These prediction markets are using RPC nodes to track the bets of users, tracks the spreads of the different markets. You can only do this via an RPC node. And when their customers are betting, those transactions are going off to a validator to validate it in the latest block. Probably the one that's most popular now is decentralized exchanges. Decentralized exchanges are checking users to make sure they have enough funds to perform the trade. And then once you submit the swap, it's broadcasting that via the RPC node as well. And so all of these applications, you definitely need a node.

So saying nodes are just an integral part of the industry is still a bit of an understatement to me. they really are the key piece. Every application in Web3 is reliant on nodes somewhere, whether they host it themselves or are using a third-party provider. The hosting needs and the cost on running all this infrastructure are exploding. There's a lot of different chains. All these nodes need to constantly be on and be in sync. There's tons of data that's coming in and out of the blockchains,

24 hours a day, 365 days a year. And so the storage requirements, the computing requirements, are growing quite heavily. These increasing hosting bills are starting to become top of mind for folks in the industry. Super helpful. And you touch a little bit on it, but what goes into running a node? What must a node operator do to spin up a node? What step software service providers are necessary today?

Most providers abstract this away. So I like unpacking it. So people understand really what's going on here at a fundamental level. First thing you need is a hosting service, a cloud provider. Most people are using AWS or GCP. These nodes are compute intensive because there's a lot of chains and a lot of data coming in and out. Some of these chains like Polygon, Arbiturum, B&B, they're growing really fast. So the storage requirements as well as hardware requirements to operate them are quite intense.

So then on top of the hosting service itself, you need additional components mostly provided by third-party service operators, Cloudflare, VMware, Kifana. and then on top of that you put the Node software, which you receive from the foundation itself. You run and manage that. Beyond just setting it up and running the software on a hosting provider, you need to definitely actively manage the updates. Chains go through hard forks. They update their versions.

That requires a good relationship with the foundation. Then you need to manage your hosting requirements on these hosting providers, which can change quite rapidly as well. RPC node infrastructure companies have obviously been around now for seven, eight years or so. What's challenging for companies' protocols that leverage these existing providers today? There seems to be obviously a number of things in the stack. Is it speed? Is it cost?

What are the big issues? Historically, companies definitely have not wanted to run the nodes themselves. I think most people outsource that to third-party providers, although we're definitely seeing they want to control more of their infrastructure, they want to control more of their costs, so they're starting to run nodes themselves. But a lot of it's been outsource, which is why we saw the growth in valuations and some of these infrastructure company providers. The largest problems that Web3 companies face with outsourcing their nodes is accuracy of the data, uptime of the node, latency between when the person asks for the information and receives the information, and then ultimately the cost. So these Web3

companies, wallets, custodians, dexes, prediction markets, etc. They have dependencies on node providers. And then the node providers themselves have dependencies on other providers in the stack, the hosting provider, the orchestration system, the software that they're running. At Nirvana, we've tried to remove as many

of these dependencies as possible by building a hosting service from scratch and an orchestration layer from the ground up where we gain more control over accuracy, uptime latency and most importantly, cost of these compute-intensive Web3 workloads, which they're only going one direction. They're only getting more intense with faster chains, and they're only getting more intense with more storage requirements. If we dive into Nirvana, walk us through the product offerings in how folks are using

you today. What's different? At the most primitive level, we're a hosting service or alternative cloud provider. We build and design our own servers. We've built a custom storage system when you're asking the blockchain for data speed is most important. So getting that data from the storage system is really important. So we've built a custom storage system on our hosting service. We've spread those servers and data centers around the world.

And then most importantly, we've built our own software orchestration stack on top of the servers to become really the most advanced Web3 hosting service in the market. We have a cloud team from companies like AT&T and AWS, which have, built and managed this hosting service. So we offer folks access to this hosting service to help on some of these key Web3 metrics, uptime, latency, and cost. Separately, we have a blockchain nodes team, which sets up and runs RPC nodes on our own cloud as a managed service for Web3 firms. So we're really at a quite unique space in the market because Web3 firms, they can come to us, they can host their own workloads on us. We have managed services on top of our cloud,

so it's similar look and feel to another cloud hosting provider, or we can manage some of the workloads for them across nodes and Kubernetes and things like that. How does Nirvana compare from an uptime and latency perspective? And what chains are you guys supporting today? Comparing to other providers, there are no explicit Web3 hosting services, which of course is our large bet as a company in this industry. I believe personally the Web3 industry is large enough to warrant a purpose-built hosting service. Our users can feel confident bringing their Web3 workloads to us that we've designed our software

orchestration stack with these heavy compute workloads in mind. They can definitely be sure we're not going to kick them off of our service like other hosting providers. And we've designed it in such a specific way to manage the cost of these really specific applications. In addition, we offer like really specific Web3 customer support. If you email some of these other hosting providers that you're having a hard time running a Polygon node, I'm not.

not sure what exactly their response would be. On the node side, we're reliant on very, very few third-party service providers, which means we can control costs much better than other node providers. In addition to pure cost management, we're really nimble on geographic locations, and then the software layer that we built is battle-tested across these node workloads to increase performance metrics like latency, uptime, and alerting once the node is actually set up. On the number of chains we support, I think it's something over 40. I'm not going to name them all, but it's both EVM and non-EVM.

People can check out our website to see what nodes we support. And then the cloud is available for people to execute any type of diverse workloads, which just require like a heavy amount of compute. This could be something like indexing. This could be something like a dex. This could be nodes, validator nodes, RPC nodes. We can handle like a diverse number of workloads on our cloud directly. How does a client typically start working with Nirvana?

Like, how long does it take to get up and running? It's super fast. I mean, clients can start provisioning managed nodes or cloud resources instantly on Nirvana, depending on location and availability. We've grown revenue something like 650, 750% this year, and we just raise more money to increase our resources due to demand. So everything can be provisioned pretty instantly. And we've got a pretty diverse set of geographic locations for people to set these workloads up in. You touched on it a little bit around the bet that you're making is that you're making is that that crypto is big enough for there to be a purpose-built cloud provider.

We've seen the rise of purpose-built AI cloud infrastructure. Why do you think we haven't seen more crypto-specific cloud infrastructure companies emerge to service the industry? The way I would sum it up is, I think the boom and bust cycles in crypto and some of the regulatory uncertainty has prevented like a proper investment in a crypto-specific hosting service. Our big bet is things have changed. people, they're running these workloads in AWS or GCP out of convenience and perhaps scale,

but it's coming at a huge cost and a huge tax to the industry. There's other super low cost providers out there as well that people have tried, and the crypto industry tried to use them as a hosting service, but they've got a history of kicking them off and they lack in performance and maybe they don't have the right geographic locations. And so our really big bet is, of course, that time is right for a crypto-specific cloud company. By the way, it's not exactly zero-sum. There are technologies which we use, our clients are using where you can actually tie clouds together

with a private connection. We have multiple customers which are effectively using Nirvana for those really intensive compute, really heavy on memory workloads, really low latency, like where they need to grab the data from the storage, super low latency workloads. They're running those workloads. in Nirvana and passing the data back and forth between Nirvana and like their AWS instance, which is where maybe they're hosting their front end or like their main application.

So cloud's definitely not like a winner take all. If you read things like Gartner Cloud Research reports, the compound growth rate of cloud computing is enormous. And last year they attributed a lot of that to crypto and Metaverse. I don't think this is a winner take all. I think in traditional businesses, people are having like a multi-cloud strategy, it just hasn't gotten there yet in crypto. And so I think the time is right, and that's our big bet. And are you finding that Nirvana clients, there are use cases that are unlocked by leveraging barramental servers? For sure. I mean, bar metal is a pretty hot topic in crypto is cheaper and more performant. I would break down our clients use cases in three distinct buckets.

And I can give examples. So there are some that are just straight up looking to reduce their infrastructure costs. There are some that are looking to increase performance metrics. And then there's a third one, which is just doing it for pure redundancy purposes. So we've got a really big firm in the U.S. that's a unicorn in the crypto space. They run 60 networks internally. they have moved some of their workloads off of Nirvana in Frankfurt and Silicon Valley to manage their costs. So for them, it was a pure cost management exercise.

We've got two high-frequency trading firms in Asia. The name of the game in high-frequency trading is latency. So they wanted to co-locate some of their nodes with some customizations in our Tokyo location to help reduce their latency. So for them, it was not so much cost, but more leveraging bare metal and a cost. custom-built hosting solution to reduce the latency required on their nodes for trading purposes. The last example, which is kind of interesting, is we have an actual node provider as well, infrastructure company using us. They like us because they can run their own nodes on our cloud, but then tap into our managed nodes that we offer for redundancy and somewhat of like an

insurance product, a backup product. All of that's available on the same platform and helps with their redundancy and backup. Do you think some of these mainstream crypto products, the Bitcoin ETF, the Ethereum ETF, does that impact the institutional node operation industry at large? How do you think that this industry evolves over time? I think it's only a matter of time before staking is approved within the investment context, whether it be ETFs or asset management. I think it's just another thing for people to charge management fees on and collect

taxes on. So I think it's going to be approved. So I think the institutional wave is coming, which means the infrastructure must be set up to handle the coming wave of this. I know we saw Galaxy acquire a staking provider, which means they're looking into this and getting involved. And so on the institutional side, I think they care about three things, most importantly. Number one is that the hosting provider has gone through the right audits. So stock audits, ISO, whatever they prefer, we're currently going through sock audits now. Then there's geographic location. They want to pick specifically where they're located at, almost down to like the data center

level, not just like a general region. And then there is the key management part. So RPC nodes don't require managing a key, but some validator nodes require managing a private key. So hosting providers, staking providers are definitely going to be grilled on how they manage their keys, what kind of service they're using to secure that with this in place. And I know a lot of companies are focusing on this. I think there's going to be a lot of M&A in this as the wheel of institutionalization.

We've just gotten to ETS, and I think staking and yield and things like that are coming next. You touched on it a little bit, but how will the landscape of service providers change in its own right? You think this is all just part of this institutional wave? I think the ones to watch are Google and Amazon. They're in like a really tough spot. They're hosting a large percentage of these nodes via the providers which are using them as a hosting service. But it seems like they're going to run more and more of these as a managed service for developers as well. So they're in a weird way sometimes competing against their clients, which are using their hosting service.

So what I've seen is a lot of the pure node providers are trying to diversify and other things upstream like above the nodes. So Alchemy's launched an indexing product and Quick Node just actually yesterday last night launched a roll-ups as a service product. You can see the node providers are moving upstream to more like value-added stuff. And then another assumption that I have is that a lot of people are going to be running their own nodes as they only kind of focus on core chains and their customer experience. It can be a huge competitive advantage to run your own nodes. There's also two other thesis that I have is true institutional firms will acquire small staking providers to bring the tech under their regulatory umbrella. and then in the validator and staking side, the only thing that matters, in my opinion, is distribution.

I think there's going to be very large crypto consumer applications. Think about applications that have like really large retail user bases. They're going to start to run nodes as well. They're going to acquire small, maybe struggling, staking companies. While the tech is important and security matters, it doesn't matter if you don't have people to like stake with you. Institutions are coming, but I still think it's largely, individual and whale based, for lack of a better term. And so I think an area to watch would be

consumer applications getting into like node running and stuff like that. But as the industry matures, I still think there's just going to be like a large abstraction of nodes and people are going to start running them themselves. It can be a huge competitive advantage to run node yourself. Validators are really interesting because whether you have a million dollars staked in the validator or $100 million stake in the validator, the cost to run it is exactly the same. So it's really a game of acquiring a lot of stake. Something to pay attention to where you can kind of separate some signal and some noise is do people run nodes?

Running an RPC node doesn't have any direct financial incentive. So do people run nodes because they can monetize it and make a margin in running that node? A lot of foundation subsidized node running because they haven't reached. to scale yet. One way to maybe gauge the health or traction of a chain would be are people running nodes without any direct financial incentive from the foundation? Do they have enough customers that are willing to pay for that data because those customers are building applications on top of the chains and are able to monetize them and cover the cost of the nodes? This has been great. Super excited to be partnered with you. I think this is incredibly informative

for the listeners. Any closing thoughts? Where can people find you? Where can people learn more? about Nirvana. We're pretty active on our Twitter and LinkedIn. So it's nirvanalabs.io, Nirvana Labs on Twitter. Happy to talk with people, engage with people on LinkedIn, telegram, whatever they want. We give people cloud credits just like Amazon and Google. So any listeners out there can say you heard us on Castle Island and we'll give you some free cloud credits and talk to you about your cloud usage. Check out the chains we support. And if any of those are interesting, we can help you guys get set up with those as well. Awesome. Thanks so much for joining me today, Dan. This is great. All right, Sean. Thanks for having me.

Thanks for listening to another episode of On the Brink with Castle Island. To find out more about Castle Island, visit castle island.Vicc. To listen to all of our podcast episodes, please go to on the brink dashpodcast.com or just click on the tab in our website. Thanks for listening.