At our Crypto Driven event, we hosted co-founder and Managing Partner Kyle Samani, who is widely recognized in the crypto ecosystem for his writing and system-level analysis.
We covered a bunch of interesting topics including:
- Multicoin’s portfolio strategy and preferred investment structure
- Their big bet on Solana
- Why is composability important
- Why they’re excited about Web3 infrastructure as a key investment them
Below is the video and full transcript.
(Crypto Driven is a team effort – many thanks to my FirstMark colleagues Jack Cohen, Karissa Domondon and Diego Guttierez)
TRANSCRIPT [edited for clarity and brevity]
[Matt Turck] I’d love to start the conversation with Ukraine – first, obviously it’s on everyone’s mind this week, but also it’s been a really interesting moment in the world of crypto with a lot of almost contradictory trends. The discussion around Bitcoin being a hedge against market turmoil, this discussion around the Russians being able to bypass sanctions by using crypto. What do you make of it all?
[Kyle Samani] It’s really unfortunate, obviously what’s happening with Ukraine and we have our best wishes out for the people there, and hopefully this all gets resolved quickly. In terms of crypto, there’s a lot of different angles to think about this. A lot of people are saying, you can use it to evade capital controls. Then that creates inherent political statements.
I think regardless of your political views, I think it’s pretty interesting that I think something like probably 30 or 40 million [dollars] has already flown into Ukraine just via donations in the last four or five days via Bitcoin, Solana and Ethereum, so that can help move the needle for a lot of people. I think it’s important to see that. I think probably the most important thing that will happen from this is just increased awareness that crypto can be a force for good in the eyes of a lot of people.
My general sense is that the vast majority of the world is empathetic to the Ukrainian position here and the position of the Ukrainian people and seeing payment rails work to get money to normal civilians is a pretty good use case for this technology. So I think increased awareness, increased adoption. I don’t think you’ll have any serious political backlash. I say overall net positive for crypto.
Do you think that’s going to accelerate regulation in the space? Christine Lagarde of the ECB just made statements in that general direction.
Yes, it probably will increase the pace of regulation in Europe specifically, I don’t follow European regulations as closely as I do American ones. The train has left the station. American politicians and bureaucrats are trying to figure out how to regulate crypto. That has been an ongoing process and will continue to be, this may accelerate that a little bit, but I’m not sure it will seriously move the needle. My sense is there were a lot of European countries that weren’t paying attention at all. And that this is now, given Ukraine is not very far away geographically is going to force the issue for a lot of them. Increasing the pace of regulation is net good for the industry as a whole because it reduces uncertainty.
You’re a VC in the crypto space. What does that mean exactly? What’s the role of a VC in crypto and what should founders expect from a crypto VC in a context where a lot of people are talking about getting financed in different ways in that space?
In a lot of ways, my job is not that dissimilar from your job. I spend most of my time reading, thinking, talking to smart people, trying to figure out what the future looks like.
And tweeting too, right? That’s part of the job [laughs].
Twitter’s a major part of the job. I’m not as witty or as funny as you. I think I get a lot more people yelling at me, than people yell at you. I guess you’re doing something right that I haven’t figured out yet. That’s all fairly standard.
I think what we do differs from most VCs is in a few respects. First is, I would say the degree of overall portfolio correlation we have and portfolio synergy is much higher than traditional venture firms.
Even for firms that are specialized, let’s say a SaaS fund or a biotech fund or a healthcare fund or whatever, or even if you’re a marketplace investment fund, the amount of correlation between your investments is fairly low. The amount of synergy between them is usually quite low. Generally speaking, we try only to invest if we have portfolio overlap and synergy. That is by definition riskier from a portfolio construction perspective, but also means we have theoretically more upside, if we’re right as a whole.
A large percentage of our time is spent helping all of our portfolio companies navigate the rest of the crypto ecosystem, figuring out who to talk to in other sectors of the space, whether that means existing portfolio companies of ours or whether that means other players. We obviously spend a lot of time coordinating and helping those entrepreneurs build relationships and we’ve got to host events for our founders to just help them get to know each other better because our hope is they will all work together in these systems.
Do you help with things like token design and the more specific parts of the crypto ecosystem?
Yeah. I would actually say the thing that I think we probably do better than just about anybody is what I would call a full stack capital market strategy.
The closest proxy to this in the traditional world is investment bankers for helping the company IPO. But we do it in a crypto native way. So if you think about a public traded company, they have earnings calls once a quarter, those are very structured. You read text, basically. And there’s a Q&A and only a select number of people are allowed to be in the Q&A.
You look at how information disseminates in crypto, it’s like the founder tweeted something and that was meaningful information or they posted something in Discord or they write a Medium post or a Mirror post or whatever. The overall mechanisms for information dissemination are much more fragmented in crypto. And that means it’s actually harder as a whole to convey what you want to the public at large so they understand what it is you’re building and why you’re building all these things.
So one of the things we do is a full stack capital market strategy. That means how to design the token economics of the system, how to think about value capture, how to think about staking rewards, inflation, all of these kinds of things. This means how to engage market makers and liquidity providers in exchanges. This includes what message to the public at large. The traditional rule of thumb in startups is the only stakeholder group that matters is customers, at least in the earlier stages.
In crypto, that’s not as monotonically true as is in traditional consumer because you have all these retail speculators out there who want to learn. They want to get educated. They want to invest and be part of the system. And so, you really need to actually cater your messaging to this different group of people. And it’s not only in English, but it’s lots of times in Chinese and Hindu and Spanish and whatever else. That entire set of processes and operations effectively doesn’t exist in any startups anywhere in the world because series A or series C stage companies don’t do those things, obviously.
And to the extent they’re done by later stage companies, it’s much, much, much more organized and actually limited quite frankly, in what you can do. So, we spend a large percentage of our time and energy on this. We actually have one of our partners at the firm, his name is John Robert Reed, and this is exclusively what he does is help portfolio companies with this range of issues.
What do you see in rounds currently in the way they’re structured and how you advise founders to structure their financing? Equity vs tokens. what syndicate do you want around the table?
The most common structure we see today in rounds is equity with a token warrant. I’d say 85 to 90% of what we do these days is that – mechanically it varies over time, but that seems to be the standard lawyers around the world are mostly agreeing on.
Just because the mechanics are equity plus a token warrant in practice, we only care about the tokens. We assume the equity’s worth nothing. We don’t really care about the equity, that’s not why we’re investing. That was part of the question, the second part of the question is round structure and composition.
We generally prefer to be leads when we invest. That’s not a strict rule, but I’d say we end up leading 85% of what we do.
Our objective after that is to pull in everyone else that we think is value accretive.
I’ll just give an example. We just led the Ceramic round that we announced a week or so ago. Ceramic is basically a crypto native database for consumer social applications. In a round construction like that, wouldn’t it be nice if you had a ton of builders in a space, a ton of people who build other web3 infrastructure like The Graph and arweave even and other kinds of things. And so, I probably personally facilitated 10 to 15 introductions personally for Ceramic. We wanted to pull those builders into the round. Obviously a lot of those people we’ve either already invested in or I already have a direct relationship with. So, when I call those folks and say, “Hey, we’re leading the Ceramic round.” They usually get pretty excited and we can pull together a good round composition. Sometimes if we’re doing something that has more, I’ll call it industry tie to it. Like a Helium, for example or Audius or something else, we obviously very much want to make space for other strategics to be involved. We’re fairly flexible on our mandate.
We don’t have strict ownership targets for, we need to own X percentage of the token or whatever. We don’t have any mandates on entry valuations. We invest in things at a $10M entry, we invest in things at an $8B entry and everything in between.
Just thinking of the Ceramic round, for example, there’s a bunch of firms that were listed as part of the round. Is that something that you’re seeing, widely syndicated type rounds, even though they have a strong lead like you? And is that what founders should be looking for?
Yeah. Again, Danny and Michael and the team at Ceramic ran a very, very structured, methodical fundraising process. They picked us and USV to lead the round, and then they wanted to just pull in as many builders as they could because they recognized, they wanted all these builders to leverage Ceramic in some way, shape or form. That was very direct on their part. At that round size of $30M, the bulk of it was us and USV, but there were still many well in the seven figures of allocation for builders in the space.
And so, that is common at that round size. Our preference is on the earliest stages when you’re talking one, two, three million dollar tickets, there’s fewer names because you want everyone to care. It’s very easy to spray off $25,000, $50,000 tickets and not care. So our preference in the earlier stages is to have one or two people, not 10, but for the later stage rounds, we’re very supportive of pulling in the ecosystem just to get the rest of the ecosystem bought in as something that starts to scale.
Multicoin is well known for having made early and very successful bets on Solana and the Solana ecosystem. Do you want to walk us through the thinking, how you identified it, why you decided to make those bets, why you’re so bullish on Solana as opposed to other L1s?
A few things that stood out to us about Solana from the early days, one was Anatoly’s background. Basically every other layer 1 founder has some pretty strong academic background in some form, and Anatoly despises academia. That really stood out to us.
Secondly, relatedly, his entire career has been spent optimizing large scale distributed systems. He’s been doing this with cellular towers, with kernels, with operating systems, with storage systems at places like Dropbox and Mesosphere and others. He’s approached the ‘make networks of computers go as fast as possible’ problem for 20 years at every angle and at every part of the stack. And that struck us as very important background to think about doing these things.
But the third and probably the most important thing that always stood out to us from the beginning was a real focus on use case. It turns out that the most important use case for blockchain is trading tokens. That is by far the most important thing. And really I would actually argue every use case of tokens is a derivative of people want to trade stuff.
If you think about trading stuff, the obviously most important question is, well, what’s the price at which two assets are being traded? The mechanism for pricing assets is well known and should it’s called the order book. If you look at the original Solana pitch deck, it was actually called Loom protocol before it was Solana. The subtitle of the title slide said Nasdaq for Blockchain or Blockchain and Nasdaq speed or something to that effect.
Obviously, they dropped that messaging a long time ago, but their use case intention from V1 was always, we want to run an order book on the system. And that struck us. Even in May of 2018, when the first round was coming together, we didn’t appreciate the magnitude of the statement “blockchains are useful for trading tokens”. I don’t think we appreciated the simplicity and clarity of that, but it struck us as this guy understands a thing he really wants to focus on. And we understood that thing to at least be pretty important, [but] we actually underestimated the importance of it.
For that combination of reasons, we started to get very excited about Solana in spring, summer of 2018. Over the next six to 12 months, we would actually continue to increase our bullish even more in our conviction, more as we started to understand Ethereum changing their scaling roadmap for the seventh time in two years, and then started to think about as Uniswap launched, as Compound launched at the end of ’18 and the early parts of ’19, we started to recognize composability and what it was going to mean. A guy named Dan Elitzer wrote a blog post around that time called Superfluid Collateral in Open Finance and he was basically highlighting composability in DeFi.
Do you want to, maybe just briefly define what composability is for anyone that doesn’t know?
Sure. And actually, if you’re curious, I gave a presentation on composability in the context of Solana a couple months ago.
The quick version of it is, the definition of composability is basically enabling one piece of state in these systems to be used in another application, without the original developer knowing anything about what the future developer would do.
Composability has existed in Web2, broadly speaking, you can just call all rest APIs, some form of composability, which is accurate. What’s different about crypto is that the state is open and permissionless in addition to the API calls and the function calls.
And by the state being composable as well, it unlocks new forms of things that weren’t really possible. DeFi being the first example. We’re now starting to see data forms of composability and networks like Ceramic. But anyways, back in early ’19, we started thinking about how are you going to make Uniswap and Compound talk to each other? It was very obvious, you wanted to do margin trading by combining those two things. And then we started thinking about scaling the number of users, and that’s when the light bulb started to go up.
Ohe light bulb went up, turned on, and we said, “The most important thing you’re going to need to preserve as you scale is composability. And the more you break the composability in the systems, the harder it’s going to get to make these systems really intelligible and usable, as you scale them.” You can take all of this all the way to the logical extreme and probably the DeFi use case that is the logical endpoint of all of composability is the notion of a prime brokerage.
In that keynote I just mentioned, I actually talked about a DeFi prime brokerage called marginfi, that we just announced an investment in maybe a week or two ago. If you think about the notion of prime brokerage, the whole point is you have positions across five, 10, 30 different venues. You want them to all cross margin and cross collateralize and net against each other. The only way to do that is with composability. And so, well, we really think about what does the future of DeFi look like. It’s going to be better than J.P Morgan and Goldman Sachs as far as prime brokerage goes, but the only way you’re going to do that is with composability.
Solana is a high performance and one blockchain in particular because it runs on GPUs. Is that correct? It was designed for massive scale, is that the right way to think about it?
Yeah. There’s a lot of innovations in Solana that help make it go fast, but the most important two are the processing layer and the networking layer. In the processing layer, they have a run time called Sealevel and the most important thing in Sealevel is parallelism. Sealevel runs natively on GPUs, modern GPUs have 10,000 cores in them and video GPS of about 10,000 cores in them these days, that’s like on a $600 graphics card, kind of a thing. And you can execute transactions in parallel across 10,000 lanes.
That’s really the most important thing because – the absolute amount of computational script that you gain that way is very large. For context, if you look at Moore’s law in the last decade, Moore’s law died for single threaded processing like 10 years ago, maybe 13 years ago. And almost all of the gains in the last decade have come from parallelism and from custom circuits, and Solana very much aligns with that. The change in the heat dissipation and silicon design.
The other major breakthrough, not actually breakthrough, is just innovation in Solana, is their networking layer and their networking protocol, called Turbine. The basic way to think about Turbine is you have data propagating between these nodes. Every node is sending one bit of data it gets to two other nodes. Those other nodes are sending the data to two other nodes etc. It’s a BitTorrent inspired data propagation system. This is also done with erasure codes so that even if you have faulty data, the data is being dropped as it gets passed along, you can still recombine the data at the end.
And so, this data structure is about as efficient as you can get for propagating data and the network through a real time network among nodes. That’s what enables you to just obviously get the transactions between everybody. And so, those two things are the things that really help you get to this kind of performance.
Let’s jump into some of the investment themes you’re excited about. Clearly, as you started alluding to through talking about Ceramic, web3 infrastructure seems to be one and also saw another interesting one that you announced recently called Fluence. Do you want to talk maybe more specifically about what Ceramic does, even though you alluded to some of it, but go into some more detail and then Fluence and why are you excited about both and the investment thesis around web infrastructure in general?
Yeah, happy to. Web3 infrastructure, I loosely think about as everything that’s not an asset ledger that is providing some either compute storage or networking function for something. I consider asset ledger, meaning Ethereum, Solana, Polygon, Arbitrum, StarkWare those things. I call them all asset ledgers, whether they’re layer 1s or layer 2s. Those have a very distinct set of technical tradeoffs in their systems and they also have very unique token designs, but then there’s a whole other class of things, Graph, Livepeer, Ceramic, Fluence, NuCypher, arweave.
There’s a whole bunch of these things that provide various technical functions out there. We’ve been investing in that sector since I guess, early 2018 with Livepeer and Graph and have continued to do more in that space. My general sense of the world is there will be, in five to 10 years time, a large class of applications that are predominantly user owned. I don’t think that necessarily means Facebook and Twitter and Telegram are displaced. I suspect those applications as they currently operate, will probably live and continue to operate in five to 10 years.
But I think there will be new classes of applications where the users understand that a primary property of the application is that Facebook or Google or Telegram cannot manipulate it. And the users are in control of the system in some way. If you believe that such applications will exist, then obviously you need infrastructure to power these things. We are probably the single most active investor across the full range of Web3 infrastructure.
I’ll then highlight Ceramic and Fluence specifically, which we announced about a week ago, maybe two weeks ago.
Ceramic is what I think Solana is to asset ledgers, I’m hoping Ceramic will be to databases for personal information. Ceramic, it’s not a blockchain. The thing that looks like a blockchain, it’s certainly a blockchain inspired data structure, but it is simply not a blockchain.
The biggest goal in Ceramic is if you’re going to enable decentralized Twitter or decentralized Reddit or Facebook, if you’re going to get to that degree of scale, we’re talking hundreds of millions of transactions per second now, you’re never going to get that all on one global ledger that comes to consensus every few hundred milliseconds. That’s just not possible. The goal there is what if we split up the state and within a sharding mechanism, and what if we actually limit specific notions of state, so you can’t have shared state?
If you think about a token, a token fundamentally is a shared state. If I have a token and transfer it to you, that means I have fewer tokens and you have more tokens, but the token itself is not… The idea of there being a token is a common good. And it’s just that you can redistribute who owns which tokens. That’s true for both fungible and non-fungible assets. Ceramic does not support notions of scarcity at all. It only supports notions of I’m writing a tweet or a Reddit message or email or whatever, but I can write an infinite number of those things and I can own them in my own data store.
And so Ceramic is a very opinionated system where the point is that it limits a lot of these things, but everything has to be owned by a user. But then the nice thing is the scale is very horizontally where you can imagine, one set of nodes has accounts zero through one million. The next set of nodes has accounts one million to two million, etc. And the system scale is very clean in this mechanism accordingly. So we lead Ceramic. Ceramic’s been around for 12, maybe 18 months is a pivot from a company called 3Box that was building this thing called 3Box IDX before.
And we’re very excited, today there’s a ton of developers in the crypto ecosystem leveraging Ceramic. People like CyberConnect and Converse Space and Satellite and Livepeer, a whole bunch of teams are building various kinds of social applications, Orbit, and a bunch of others that leverage Ceramic. I wouldn’t say there’s been any major breakouts yet, but if you’re following the space, what Ethereum felt like in 2016 is what Ceramic feels like today. So we’re super excited with Ceramic.
The other one we announced about two weeks ago is called Fluence. Fluence is arguably the most ambitious technical thing we’ve ever invested in. Certainly one of the most ambitious. The simplest mental model for Fluence is permissionless, decentralized, AWS Lambda. The idea is you could take any public piece of data, whether it’s from Web3 or from a Web2 API call, you can issue a transaction to the Fluence network and say, “Go compute on this piece of data, execute some of the deterministic computation, and then deposit the output of this computation to some data source.”
Again, whether it’s a Web3 store or a Web2 store, and then developers can specify how much redundancy you need, latency and response times, regional lock, regional differences if you need to do or not do computations in certain places. So you can specify all of this, but it’s a really cool idea is that you can theoretically scale AWS Lambda functions across a global network of nodes that don’t trust each other to execute any arbitrary computations or arbitrary function calls on public data.
I realized that was a mouthful of abstract words strung together, but we certainly are very excited about Fluence. What is this useful for is actually another key question. Some examples are snapshot votes and governance votes are very useful to compute and aggregate in this model. I expect to see a lot of on-chain games start to leverage this for various types of function calls between them, where if the developers of the games don’t want the central server to manage everything, you need to have servers that have compute resources that are not local to the user.
I think they’ll start to see a lot of game centric function calls live on Fluence. I think over time, I’m expecting to see some really cool things emerge, like ML models, for example, if you’re going to have decentralized Twitter where you’re going to run the ML so that you don’t have a strictly chronological feed, but you have some sort of intelligence. As you have more and more data in Ceramic, as more and more data goes in the arweave, that naturally increases the interesting space to compute over that data. And our hope is that a lot of that will run over the Fluence network.
Another them you’re excited about is the reinvention of the music industry through crypto and decentralization. You wrote a very interesting blog post recently. Do you want to talk about the high level and maybe people can read the whole thing on the Multicoin website, but a high level thesis?
I published a blog post on Friday last week called Rebundling the Audio Value Stack or Chain. The basic theory is the internet unbundled the record label. It took the distribution function and broke it off. By breaking off distribution from the rest of the record label function, it’s created a lot of inherent conflict in the record label business model. With the growth of Napster and then Apple and then iTunes and then Spotify, that conflict has gotten worse over time.
Our operating theory is that you can recombine the record label value stack in a decentralized way where fans have economic interest and rights in the overall system as a whole. And that by introducing this new vector of ownership and revenue sharing and profit share, you can actually redesign the entire stack from risk sharing, to marketing and fan engagement, and then finally distribution. We’re super excited about this rebundle. Music is without question, the most important piece of creative outlet in terms of culture across basically all cultures.
It’s true in English, and it’s true in Hindu, it’s true in Chinese as well. And it’s the most under monetized compared to its cultural impact. We’re super excited about the opportunity to change notions of ownership and monetization models of music. We think that the logical net conclusion of all of this is what we like to call a music VC DAO, all of which is referenced in my post that was published on Friday.
A few questions from the group, and we’re going to jump back and forth between some of the topics we covered during this conversation. Question from Andrew, “How many builders and developers are working on Solana worldwide, and what has been the incremental growth since the Wormhole hacker event, and how do you see security as a future issue to rectify these attacker attacks?” The growth of developers and security for Solana.
I don’t keep track of developer metrics with any closeness. The thing I optimize for or I should say, look for is developers I’m willing to fund. I realize that’s not useful data to most people, it’s turned out to be very useful data for me. And it’s obviously my full time job to try and answer that question. I can tell you, we have written in the last six months, invested in 20 to 25 Solana based teams and we’ve passed on probably 50 to a hundred more that have been funded by other people.
There’s hundreds of teams that are not venture backed building on Solana. And that’s really the thing to look for. I would look at the number of people who can raise outside money is probably the best proxy and there’s data on that. BlockPublisher and stuff, and other people do as well. I reference numbers in my keynote presentation that’s linked somewhere in the chat. There’s that data. The second question is security and the Wormhole hack and stuff.
My general view of bugs and software is they don’t matter in the long term. They’re obviously very painful in the short term, but the general rule of thumb of software is if you can create anything in software that you can think in your mind. That statement must be true because that is the fundamental premise of software. And so, as long as you’re not trying to solve a logically intractable problem, the problem is tractable. Bugs are bugs. It’s unfortunate that people lost a lot of money. In the case of Wormhole, Jump can pay $300 million and they did, it was a painful learning lesson, but not the worst thing in the world. I generally don’t spend any time thinking about bugs. In the long run, these bugs don’t matter.
Question from Arjun, “How do you think about the timing of hopeful crypto projects generating revenue or value external to the Web3 world?” How quickly do you expect, my words, not his, but how quickly do you expect the project to become a real world in terms of generating cash or value?
I think the answer is not anytime soon, for the most part. The instructor example here’s the internet. And it turns out the substantial majority of early internet companies were primarily just people selling other internet things to each other. There were a couple of notable exceptions, eBay and Amazon probably being the most important, but for the most part, most of the other early internet success stories were softwarey things for softwarey people. Marc Andreessen wrote, “Software’s eating in the world” in 2011.
And even in 2011, the iPhone was three or four years old. And it really took probably another three or four years after that before the iPhone started penetrating every single vertical, by pizza delivery and whatever guys and radio towers using iPhones for documenting their work and all these other things. It turns out forecasting timing is very hard, when do you get the real world impact for these things. So I have no real conviction on particular timing or how to measure the real world, but I think that’s actually the wrong question.
I think the right question is what are the new kinds of things emerging that are fairly circular and crypto-y for crypto-y people and are those things interesting? And the answer to that question to me is, overwhelmingly yes. NFTs are certainly very interesting. DeFi is very interesting, creator monetization is very interesting. Helium is super interesting. And so, a lot of these things just have to be circular for a little while before they break into the real world.
Most people are trained to believe that circularity is a bad thing, because it oftentimes is how you get to MLMs and Ponzi schemes and other things, which is true, but it turns out most of the biggest breakthroughs in the world required just a lot of circular belief and investment from a lot of people for a long time before they became viable. And so, I don’t mind the circularity. As long as you see the fundamental progress continuing to improve with some view of what’s the light at the end of the tunnel, then you’re all good.
Great. Well, this is super interesting. This feels like a good place to stop this conversation. But look, you have such an incredible view of the space that I hope we can do this again from time to time, you can come back and check in. It feels like we could chat for another two or three hours at least. This is great. Thank you so much. Really appreciate it.
Hey Matt, thank you very much, everyone and my DMs are open on Twitter. If you want to DM me there, feel free to.