Logan Jastremski: Solana vs Hyperliquid - Who Wins The Global Exchange Race?

You're now plugged in to the Deli podcast. Hey everyone, I'm Tommy. Welcome back to another Deli podcast episode. Today I'm in person with Logan who runs Frictionless, an incredible fund down here in Miami. Logan's been on the podcast before.

Was it two years ago?

About two years ago. We talk about ETH and decentralization and low throughput versus high throughput. Was it the worst podcast appearance of your life?

No, it was good. I think it was very factual if you look back at it now. I feel like a lot of that would still be kind of like evergreen content, which is hopefully the goal of how we approach investing. If our thesis changes a lot, then we're probably pretty wrong.

I don't remember too much of that episode to be honest. I remember it being a lot of fun, but I wonder what we got wrong, what we got right. I should have watched it. We should just post that episode and see if people pick up on the differences.

I remember distinctly though we talked about Ethereum L2s, why L2s in my opinion were not going to win, why you needed high throughput, why decentralization was objective and measurable, and that kind of the fears were overblown about high throughput blockchains not being decentralized was kind of the main things that I remember.

So when we spoke then you were full in on fund one.

Yep. Now you're moving on to fund two. We talked a little bit about your thesis for fund one versus what you want to do with fund two. What's the difference in thesis between frictionless one and frictionless two?

It's a good question. I think we got a lot of things correct in terms of our understanding that blockchains really needed to scale. And so understanding that you had to transition from low throughput blockchains like Ethereum that were amazing in terms of turn completeness and pushing the industry forward but maybe optimized for the incorrect things.

I think we got that fairly directionally correct. I think we also kind of, if my memory serves me or if I just knowing myself, we probably touched upon the modular versus monolithic or definitely did and I'm assuming my point there was that the monolithic chains were the correct architecture design and that the modular design was going to be incorrect.

In large part a lot of the things that we probably talked about a couple years ago I think has panned out. I think the thing that we've gotten much more of an appreciation for as fund one continued was that throughput definitely mattered in terms of you need scale for application engineers to build more expressive applications.

But the thing that we have really come to appreciate now is that throughput only mattered in a sense that it enabled scale. But the other part of that for at least the underlying infrastructure was execution and trading.

I would say our fund one portfolio definitely has some trading related portfolio companies within there. But I would say for fund two we've definitely have a full appreciation now for the things that make revenue and where is the value going to accrue and just doubling down on our current thesis. But in large part, it's more of a continuation than like a pivot by any means.

So, just to recap, with fund one, your thesis was high throughput monolithic chains will win, pressure on ETH L2s, the modular thesis. And now with fund two, you want to stick with the high throughput thesis, but you want by extension like the second order, you want to focus on apps and projects that lean into that and make real revenue based off trading and execution.

Yeah. So we actually plotted this and we showed Raj and Anatoli this in Abu Dhabi and they're like why haven't we seen this? But it was a chart from essentially the time from an application to get to 100 million in revenue from 0 to 100 million and how many days it took.

A lot of those companies are now currently being built on Solana and this is not just crypto companies this is across AI as well. The fastest 0 to $100 million revenue companies in the world are crypto companies. And this has been stolen and remade multiple times. I'm kind of pissed off about it, but it is what it is, I guess.

Wait. That chart on Twitter where it's zero to 100 million revenue and the amount of time you see the the lines created. You did?

Yeah. I've seen it copied a lot.

Yeah, I know. I've told I've been told that we need to be more public about our research because everybody takes it, which I guess is bullish. If I don't know if you remember, but one of the first things that we published was essentially the throughput of different blockchains.

At the time people were like, "You guys are idiots." I don't know if I concuss on this, but they're like, "You guys are idiots." Like this is not true. And literally all we did was go to each individual's website and it was just the block times and how big the blocks were. It was like super basic.

We just said this is the amount of kilobytes or megabytes of throughput per second. Looking back like it kind of made us second guess ourselves, but then like a year later, Paradigm published Gas Per Second. I was like, this is literally what we've been saying for a long time. And because Paradigm wrote it, everybody thinks they're geniuses.

I mean, they're a pretty smart crew.

Yeah, I get why you're annoyed. I mean I mean they are smart, but more broadly it's just like I think one thing that I've come to appreciate is like we definitely have a lot of alpha. We're not always the most public about it, but I want that to change because I mean the good thing is you're acting on it though, like you're actually investing based off of it.

Yeah. And that's that's kind of been my northstar for the fund. Like I think in large part I don't know whether it's the crypto industry or just venture or even like kind of money managers more broadly. I think sometimes you can get sucked into playing the wrong games which is like Twitter clout.

For better or worse we've kind of just said like that's not the game we particularly care about winning. I do think it's helpful for fundraising and potentially like seeing deals, but our northstar has always been returning capital to our investors. And so we've definitely acted upon it which has been beneficial to us.

I feel kind of slighted. I'm pretty active on Twitter.

No, no, no. But but it works. I I I'm just mess with it. There's different games. The chart of of the Solana the applications on Solana doing first to 100 million in revenue. Yeah. Like so you're talking out of all the applications built around the world, regardless of web 2 or web 3, a lot of those companies are built directly on Solana.

Is it like one company? Is it the top like a percentage of the top 10? Like how do we visualize this because we're just talking.

Maybe our wonderful editors can throw up the chart. I can send it to you. Some of them are on Solana. I mean the Axim pump Fun. Yep. Axiom Pump Fund. One of our portfolio companies, Backpack, they're not really built on Solana. Ramp was on there. I think Lovable Cursor Hyperliquid was number one.

I think it was 89 days for them to get to 100 million in revenue. So I mean so in broad part like now what we're seeing is that these applications are making money and every single one that is generating this cash flow is trading related on the blockchain side.

Interesting. And so for us I think it's kind of given us more of a wholesome appreciation for that trading aspect. The blockchains I think why kind of the high throughput blockchains have won in my mind are continuing to win is because that the additional throughput enabled essentially more transactions per second would ultimately made it better for trading assets.

I think the thing that we've come to appreciate though is that throughput by itself maybe picking on Celestia just for a second. We've seen that throughput by itself does not necessarily equate to value capture. Celestia is killing it on the throughput slide. When we initially plotted it in our data propagation chart a while ago, they were relatively small and I think by now they're the leaders by far.

But the thing that they haven't really focused on was the execution piece. And so the thing that we've seen is that the value or revenue is ultimately determined by the execution.

I would say one thing that has been really interesting within the blockchain space really for the last 5 years has been the kind of design exploration of all the ways to scale. But if you say execution is the north star and trading is the thing you need to optimize for, I think that very narrowly narrows down kind of the possible different design decisions that you need to make.

So there was like L2s which is a single server. There's app chains which could be decentralized or could be a single server. You had the modular road map. You had charting. You had like application specific chains with or like what avalanche was doing with like different chains.

I think in large part if trading is the northstar it's kind of helped clarify and our mind why kind of the high throughput chains have become more valuable.

probably a basic question, but like why has Solana won every trading and execution enabled app that's on that first 100 mil rev list other than I think you mentioned Hyperliquid?

Happy to talk about Hyperliquid too, but like why have why didn't they choose like why didn't Backpack choose an L2? Why didn't Axiom choose another chain? Like why have they all landed on Solana outside of Hyperliquid which we'll talk about?

Backpack is different just because they're a fully centralized exchange and went the regulated route, but they're fans of Solana. I would say Solana took upon the hard engineering problems. I think one of the things that we touched upon in the last podcast was talking about single threaded virtual machines versus parallel.

Parallel execution was just enabling modern hardware with many cores to do many things simultaneously. And this is something that Ethereum and the kind of adjacent EVM chains just got wrong from the day one. And so I think because they took a different architecture path, it was very hard to bootstrap, but because they were also the first mover and kind of this parallel execution, it just enabled more of a white space for engineers to build more expressive applications.

I mean, if you recall, Solana kind of got its original footing from NFTs. And that was because Ethereum NFTs were just way too expensive. The gas wars were nuts.

Yeah, it was crazy. I personally paid like 500 to a,000 bucks on a unis swap transfer. And so in 2021 when NFTs were going on, the joke was Solana was for the poor. And it was because it cost a fraction of a penny to trade NFTs on Solana, but if you're rich, and could spend hundreds of dollars, you could do it on Ethereum.

I mean the what was the Ethereum the god I can't remember the name the one where you would compete in a transaction to create the generative art onchain in the transaction.

Oh there is a bunch the biggest one it it was like thousands of dollars they would mint out of the thousand and yeah I I lost money off NFTs so I try to block it out of my

Yeah it's it's a tough road.

Yeah. That is pretty interesting. So so because of technical reasons they've all landed specifically on Solana.

I think so. I mean, so I think because transactions were cheap, users started to congregate there, similar to what happened with Ethereum and like DeFi summer, like people just liked experimenting. But it was way cheaper to run those experiments and for a user to kind of onboard. Transactions were a lot faster. They're a lot cheaper. It just felt more modern.

So like a super basic question for you would be like an Ethereum L2 offers such theoretically cheap transactions. Yep. So like if you're just deciding based on cost like wouldn't you go there because there's nothing else built or would you or there's less built you're competing with less people on a new L2 for block space?

I don't know. you kind of run into the same problems as what we probably talked in our last podcast. Most of the L2s historically were also single threaded. Yeah. And so you didn't have these localized fees markets that parallel processing chains enabled.

Some of the more modern L2s tried to do parallel processing as well which would isolate fees for specific applications or specific smart contract apps which was helpful. But I don't think I'm actually today I'm pretty bearish decentralization. More so in the sense that I think we going back we ran all these experiments and they were fine.

I don't think people necessarily pay for decentralization. I think if people valued Ethereum like say Solana had a thousand nodes and Ethereum had 10,000 like if if people really value that they would pay 10x the amount of costs in my mind. And in a large part I don't think we've seen that.

I think people care about decentralization only if it makes better product. But decentralization for decentralization I don't think people really care about. And so, to answer your L2 question, I think L2s are fine. I think they have just generally failed to scale.

There is one really cool website, I forget what it's called, but it literally measures the throughput of all the L2s. If you aggregate all of them, I think they're still less than what like a single hydrupo blockchain like Solana is doing. So, I still think there's some scale limitations. But that may be changing with Mega ETH and a couple others.

Generally I just think the user base was there just from more historical times. The assets that people wanted to trade were there and so that kind of just stuck around and now it's about continuing to scale them.

I'm not a developer. Well, I'm probably a pro vibe coder at this point, but I do feel like there's just a very low cognitive load to picking Solana. Like, no, there is. Yeah. Like it's just like if I want to build a performant app, I go there. I kind of know things will be taken care of from an inference perspective.

So I want to talk about the other routes competition wise. We could either talk about hyperliquid in Solana or we could talk about like traditional markets going 24/7 ice trading on chain or their permission chain and going from there. I think it's all kind of the same in my mind.

They're all racing in my mind or what they should be racing for is to build the global exchange. And so whether it's the New York Stock Exchange or an L2 or Hyperliquid, I kind of view them all in a similar vein. Where you have a single server located in some jurisdiction in the world, say for example purposes, they're all in New York.

If you push it to kind of the logical extremes if these things actually work and you're doing call it trillions of daily trading volume people what we have seen in Treadfi want really advantages as much as possible and so in Treadfi high frequency trading emerge because people wanted better and quicker access to information and also trades and so you colllocate next to the actual exchange or the server in instance would be New York.

So that you can receive information faster and make better trades and have more information with your models. When I think about kind of whether it's kind of a more modern version of what exists today or a single server like Mega ETH or something like Hyperliquid that's relatively centralized in terms of its node set on like the same data center or geographic region.

To me I think you can build a great exchange product but I think that's kind of missing the point. If you actually could reimagine this and build it from the ground up for in a global exchange, I think really in my mind the things that you would want is build this credible neutrality.

This is where I think going back to my earlier point like kind of bearish on decentralization for the decentralization. I think here decentralization actually matters because decentralization makes the product better where it's not bound by a single jurisdiction and everybody in the world can have kind of equal and fair access to information.

Can we walk through that a little bit? So like hundreds of countries around the world and you're saying that decentralization makes the product better. What does that mean? Like if I'm in the US versus Russia versus China using the exchange, like why does that make the product better?

So, let's use the kind of like centralized exchange example. If you're in Russia and the server for the exchange is located in New York, why would I, as a Russian person want to trade on a server that's located many miles away from me? If I'm running a quant shop and you'll always get front run versus the guys in the like it's purely a latency like physics perspective.

Your disadvantage if it's not in your home jurisdiction. And so I think even if the technology was better and this is the same for mega ether or hyperliquid like hyperlids in Tokyo like if I'm a United States person and everybody in Tokyo is getting better prices than me like if these things actually work out like this is not like today like today it's mostly retail it's fine.

If these things actually scale to where you're doing trillions of dollars a day and this is the everything exchange, I just have a hard time seeing the United States, China, whoever, like big global powers that actually trade a lot saying, "Yes, I'm okay with that data center being in Japan or I'm okay with that being in New York because I'm disadvantaged purely from a physics standpoint."

Interesting. Okay. So, I thought you were going to take it from a regulatory and like access view. So there there is that too which I think like the fully decentralized worked but this is just purely on like hey I want to make trades and I want a fair shot so to speak at making those trades.

Okay I totally understand. So if you're around the world in any country you want inclusion in a market that's in an exchange that's close to you. So you're not front run by other people that are physically close to the exchange. And so this is I mean why I don't think you have a global exchange today and why you have regional exchanges.

There exists the New York Stock Exchange, the London Stock Exchange, India, Japan etc. because these I'd say physics limitations. Now you have what has emerged is you have highfrequency trading firms around the world that have these private fiber cables that when both of these exchanges are up and running will do arbitragees in between them to bring the prices back in line which is I would argue a benefit to each of those exchanges to have a global price.

But they're doing that on the back end and it's they're connected to these exchange venues through the microwave towers and the fiber cables or even now with Starlink. But these are separate exchange venues around the world. And I would argue that they spun up not only from the regulatory standpoint, but also because of the physics and the limitations of kind of the speed of light.

There's still somebody in between those two exchanges though that that has to even out the price, right? Or up the prices back to

Yeah. And if the exchanges are up at the same time, they do arbs.

Okay, so let's continue with this line of thinking. So we're talking we want a global exchange. We want nodes in each market so that nobody's front run. So the question is who achieves that goal? And I think we could talk through three options that you brought up which is New York Stock Exchange, Solana, Hyperliquid. Maybe we could talk about how they're each moving in to achieve that goal and who you think's going to win.

When I think about the global exchange, really want you want like almost equal and fair access to that data. So just being able to read that data. Then you generally from the physics standpoint don't want to be at a disadvantage to really any other party around the world.

A very clever design that has kind of been thrown about that is just now starting to be implemented and is I would say I think it's live in suite it may be live in apttos and it's not live in Solana yet is this idea of multiple leaders in blockchains historically you have one leader and what I mean by this is generally at any given point in time you send all your transactions to this one leader so going back if we continue with the New York example the blockchain leader is in New York you send all the transactions to the leader in New York or if you're in Singapore, you still send your transactions to New York.

Those are aggregated then sorted and then disseminated. So it doesn't really matter where you are in the world. You're always going to like a centralized venue send all the transactions in in the world to a specific spot. But the clever idea that has now being started to implement and I think is going to be as necessary as going essentially from low throughput to high throughput is this idea of multiple leaders.

In the most basic terms is you can have leaders in various jurisdictions around the world where you can cut down on the time to inclusion. So, for example, instead of having someone always has kind of an information advantage, you can disadvantage everybody to call it 20 milliseconds.

So, what happens is you run parallel auctions. I think 20 milliseconds would probably be good. But under 100 is probably fine enough, too. But I think we'll get to like 20 milliseconds. You aggregate kind of all these transactions around the world.

It's interesting because from a physics standpoint when you get to like call it 20 milliseconds you can only see the transactions from a certain part of the world because light cannot physically travel to other parts of the world from like a fiber optic cable standpoint. So, for example, if you had one leader in Australia, you would probably only see every 20 milliseconds only transactions from Australia because if it's 30 milliseconds away, you're not going to be included in that block and you're going to have too late of data.

It avant it it makes it advantageous obviously for Australia because they're able to kind of like update their worldview in kind of 28 millisecond increments. But if you have these world leaders in call it Japan, United States, Hong Kong, Russia, UK, wherever the thought process is you have these 20 millisecond batches that are happening simultaneously and continuously and about 100 to 150 milliseconds later each of those transactions are essentially sorted and then finalized.

So you don't because of the speed of light around the world the world it's like 100 to 150 milliseconds it takes a while to actually sort those transactions but at any given time in 20 millisecond increments everybody you can almost think of it as they're voting on what they believe the price to be around the world so all the jurisdictions around the world every 20 milliseconds they're saying I'm buying and selling at this different price and then 150 milliseconds they're slotted.

But this entire process is continuous both on the so you don't wait for these auctions to end. You just get the new price every 50 milliseconds but people are continuously building these blocks every 20 milliseconds. That's awesome. So everybody's kind of disadvantage at the same but they're aggregated and finalized every 150 milliseconds. That's so interesting.

Does that make sense or is it

No, it definitely it's a little complicated. No, it I'm I'm getting most of it. So I kind of want to understand the impact though. So right now if we're up late trading on Jupiter or whatever on Solana and we're making a trade, we're selling soul, we're buying some random memecoin and the current leader is in New York, you and I are both in Miami.

So we're not really competing. But if a third person in Tokyo wants to buy and sell the same token at the same exact time, we clearly get filled before they do. So the question is we the question I think or the thesis is we want a global exchange that's fair to everyone around the world because we're not just going to be trading meme coins.

We're going to be deriving the price and flows and trades for every asset around the globe for all assets. The idea is so right now block times are not fast enough for this to actually be useful because Solana has block times of 400 milliseconds. Apttos is actually the fastest at I believe 50 or 60 milliseconds. So their block times actually matter where in the world you the leader is, but for 400 milliseconds there's enough of a global lag so to speak that it doesn't matter.

Got it. So everybody's transactions around the world have time to make it to the leader.

Yes. And this is why historically Ethereum has like 12 second blocks because they've mandated that you need low hardware requirements and low throughput and so they've just given everybody extra time to aggregate their transactions to this leader.

It's such a it's such a basic concept the way you describe it that we want time to include everybody. I never thought that that would be the limiting factor. I thought we would always just want the fastest possible block times. And so now the idea is if you shrink it to less like 60 to 70 milliseconds is like half of the world's circumference.

Once you shrink it to that you that's where you can start getting regional information. So the thought process is once you have multiple leaders and Anatoli has always used this example. It's kind of dumb but it's funny. There's a call it a boat in Singapore that has $10 billion worth of iPhones and it sinks.

The people that will be able to kind of capture that ARB so to speak are the people in Singapore because they are able to monetize this information first. So in this kind of 20 millisecond world block view the people and traders that are seeing this boat go down could just start market dumping and the people on the other sides of the world don't know this is happening but these trades are now just market dumping the price and they're able to capitalize on that alpha first.

It's a great example. And so yeah, so it's it's taking the aggregate worldview price and then kind of like finalizing it on 150 to yeah 150 millisecond kind of like time time stamps.

Can't you cover the world enough if you go 50 milliseconds and have three or four nodes?

Yeah, I mean, we I think like we'll definitely experiment with like do you need two like because then you just cut down the world latency in half. Like but I always use the example of like Amazon Web Services. They have like data centers around the world because they want to cut down on the latency for you to ping the server.

The more servers that there exists in the world, the lower it takes for you to ping that server and get a response. So they have Amazon East for people on the east coast, Amazon West. You just ping the server that's closest to you.

It feels like I mean I'd love your view on this but it feels like we're getting to a world where obviously we have high frequency trading shops, we have hedge funds, we have us retail traders and and institutions and funds but it seems like I mean I know this is like a loaded term but like AI agents will transact so fast in such low cost and it's such high speeds that they would really favor regional markets.

They I mean I I think whether it's AI agents or humans no one can beat physics which is why I think this is a cool design because the goal and it's my pin tweet I don't think people understand it but it says the goal for blockchains is be to become the leading indicator for price discovery.

Why I think this multi-leer world view has the potential to become the leading indicator for price discovery is because it's aggregating in 20 millisecond increments the world's average price or the world's price for any given asset. If this is the pricing indicator for the world, what will happen is the blockchain will be the true price and exchanges on the centralized venue will be the lagging price.

What happens now today is that blockchains because they're not performant and not efficient enough is that they're always lagging like the New York Stock Exchange or for example Binance. And so people view the market making there and the true price and these centralized venues. But if you invert that and you can flip it to this global kind of trading engine, so to speak, this is now the world's most true price because it's representative of essentially the global worldview.

I think that's what everybody's trying to build in my opinion. That's the grand prize. So maybe to like wrap it back up, the reason why I think like L2s like Hyperliquid and kind of like the old New York Stock Exchange aren't going to win this global exchange is because the design and engineering to build this is much different than just building a great product.

Hyperliquid has built a great product. they deserve all the praise in the world for doing that. They're making a lot of money. But it is a different engineering design to build this global exchange. I think purely from a fairness so so to speak I don't think the world and its limit will say hyperliquid is this global exchange that I feel like as a US person or anybody kind of outside of that market will trade on.

I could be wrong but this is kind of our broader thesis. If if we're wrong it's because the world did not care about these properties and that could be true. But this is our assumption and how we're kind of investing.

I think that's the clearest explanation on why decentralization or distribution is a benefit, not a cost.

I mean, it's literally like Amazon Web Service. They're globally distributed because they cut down on latency. The only reason why I think it's a bonus to have the decentralization is because it cuts down on latency for time to inclusion for the global market. It doesn't advantage any single specific person.

I want to ask you two quick questions and then go back to your your main line of thought like in a world where we have localized markets and nodes around the world. We have all these venues. It's all feeding into one global exchange. I'm trying to think through like what the benefit of that is to the world.

I'm thinking through your example the iPhones like like the old days our like equity research analysts would give quarterly estimates they would say how Apple is doing we would you know talk about you know how they're doing how we think the price should be the price will go up the price would go down in your example it seems like if people are able to effectively act on individual bits of information in a bottoms up world like the iPhones falling off of a ship leads to less iPhones sold leads to you know lower numbers less revenue yada yada if people are able to effectuate those trades in micro markets around the world.

Maybe it's on a prediction market which bubbles up to the actual price, things like that. It seems like our world becomes much more efficient from a pricing and liquidity standpoint. I I'm not sure if that's the right way to take it. I I think it's I view it as if you could rebuild finance, modern finance system, the world is very clearly global whether we like it or not.

I think having 247 markets trading all the time and having kind of an aggregate worldview it's just a better product. But but this isn't just 24 like the New York Stock Exchange could offer 247 markets.

Well, your thesis what you're describing is like efficient 247 markets. Yeah, it it it's should be in the way that I believe it will work is that it doesn't advantage any single jurisdiction and it allows everybody in the world to participate.

I think like the actually the funny enough the re original reason why I got into blockchain was Ethereum's banking the unbanked. Ethereum got me into crypto too and like I thought that was super noble because it was like all right you don't have a bank account you can't invest in assets because you don't have access to those assets.

When I think about the world and like modernizing kind of our technology stack, what it would look like, I think it would be this 247 global market that gives everybody equal and fair access. In addition to, I mean, whether you want to earn yield in stable coins or just get stable coins like trade your currency from one currency to another, buy Bitcoin, invest, save, payments, like all that.

I think if we're all kind of I I describe it as like a single API, you opt into this API. This is the global API for finance. That just helps everybody coordinate payments, value transfer, and price discovery. That's awesome.

It is it is wild meeting people that have never had to use like Fidelity, Schwab, Erade, whatever, and have only ever used crypto. Because when you talk to them about like the frictions and like how cool it is to be the frictions of the old world, they don't even get it.

Yeah. And I I mean, we were talking with some LPs and they're like, "Look, I I can touch and feel the AI like I can take Tesla FSD and like it drives me around." Like it's hard to touch and feel crypto. And I think that's fair. Most people especially in the western world have access to these systems today like our currency is not necessarily hyperinflating yet.

We have access to kind of like investment tools and generally we can move our money around. I think that's fair and most people are like if you say oh blockchains are trending towards high frequency trading their eyes kind of roll over too and they're like I give two shits about high frequency trading which I think is fair.

To me like my co-founder Diego, he had this talk at our summit and we were talking about it's not the next like version of the internet. It's not web two and web 3. It's the next iteration of finance. So I think most people in the the the United States and Western world just don't appreciate the how great we have it. And so the impacts of kind of this more modern technology are not as felt.

It's also just like it it's crazy how just such a basic education miss it is too, right? Like I've met a lot of people that have a lot of money in their checking account earning no interest. Like very basic thing that they can get with stable coins.

I want to circle back to your your global exchange thesis though. I think you've accurately laid out an incredibly succinct