Business of Block Space Pt. 2: Mevconomics | Will Nuelle - Galaxy

[Music] hello hello okay great um I think we just need the slides here we go great um hey everybody I'm will um I lead investing for Galaxy digital um we are you know new york-based Financial Services firm in the crypto space uh you know we we invest and and trade all across the industry but I lead uh co-lead The Venture efforts at Galaxy this is a thread of research that I've been pursuing and continually analyzing for the past nine months or so um today I'll do a bit of a presentation that I've that I that I've given previously and then a little bit of additional work on me um and we'll we'll broadly talk about block space as a business model where consumers are purchasing block space if you think about it as a product the cost to serve that product to block space and then the impact of me which is kind of an endo uh endogenous transaction type uh inside of this block space economy um and I want to maybe frame it from a lens that that I work out of um new technology paradigms fundamentally create new business models and this is as an investor one of the things that I uh have to be keenly aware of so there's a firm uh that that some of you may be familiar with called Insight Partners Insight Partners was formed in the mid 90s when the PC boom was basically halfway off up its Des curve right PCS were getting shipped to every home computer and insight uh to every home and insight Partners identified that SAS as a business model had a propensity a high propensity for recurring Revenue high gross margins and uh basically costs uh in R&D that scaled logarithmically with the revenue so you could really gain as you scale over time and they identified that this is a business model that was really an elite business model they built a firm around it and they were absolutely right um I think that you know that is something we are ultimately at Galaxy want to focus on in crypto what are the new business models enabled by the technology so I see four business models in our industry today which have a you know recurring sort of robust product Market fit the first is um software businesses so these are the SAS businesses that I just talked about um this isn't a new business model per se but certainly businesses like fireblocks Alchemy um and others have have sort of repeatable robust product Market fit the second is exchanges the propensity to trade uh assets of all types whether that's you know um you know cryptocurrencies or nfts and of overtime We Believe at Galaxy gen broad Financial assets um there's a there's a clear desire to do that in crypto and exchanges generate robust sort of repeatable product Market fit the third is stable coins and lending businesses what some people call net interest margin businesses which um you could think across a whole Spectrum there there's probably a whole other presentation to be done on that business model in the industry the the fourth is block space which is the area I've been focusing on and the area ultimately that I find most interesting blocks space you know it's a product that blockchains sell to Consumers right consumers who are you know transacting on a chain buy blocks space from a blockchain which produces it um what is block space at the core it's a set of operations on a global State machine where there's sort of global consensus over that over those State changes um and it's priced on a per operation basis at at a at a virtual machine level okay this is where it gets a little bit more interesting um if you look at different blockchains and the actual amount of you know let's call it block space they sell to the cons to to end consumers um there's real differences between different environments right you can see ethereum sells on the order of $10 million of block space a day um and other chains like binance arbitrum uh as well as optimism sell on the order of $100,000 a day in Block space and others you know other others less our proposition is that the difference between those uh environments is determined by the network effect that those environments have right there is a a real Network effect around this this compute um W which comes in the form of capital in users in application developers liquidity in smart contracts and ultimately smart contracts that are deployed on a particular chain um and I think that Network effect is hard to overcome block space is a commodity product but you if you have a network effect you can obtain ultimately a higher a higher margin and and a comparatively higher uh fee per transaction and so this is a way we like to analyze block space if you just look at the history of what users have paid to different environments it's actually pretty interesting right um this is through 2023 these are actualized numbers so 2021 consumers paid 10.9 billion to use block space across these six chains ethereum binance polygon Avalanche arbitrum and optimism there's others I would probably add here salana would be great salana is a fairly difficult uh chain to get dat to you know produce data on because they have so many transactions um but there were 10.9 billion in 2021 this year uh in 20120 or last year in 2023 um users paid 2.8 billion across all those chains 2.4 of that is for ethereum specifically the other things I would highlight uh on this on this sheet are uh you could see in 2022 if people remember in the market the alt L1 rotation so that's you know users moving into polygon and Avalanche you can see an uptick in transaction fee volume paid in those environments where you have you know a Avalanche and polygon growing 130 or 100% year-over-year from the prior year and then in 2023 you can clearly see a um an L2 rotation where you have optimism and arbitrum growing their fee Revenue by 100% year-over-year um the other thing I like to think about and pay attention to are what are what are what is the market paying per dollar of block space Revenue um you can see this is a a chart of it's a bit of a confusing chart but it's the fee the multiple paid on the annualized uh transaction fees of the last 30 days so at the bottom here you can see ethereum uh which the market pays about 100 times the uh annualized um transaction fees for ethereum and then you could see on the on the top here arbit uh or sorry um uh Avalanche where the Market's willing to pay up to at times 500 500 plus um uh times multiple uh which is which is which is quite interesting you also have arbit and optimism kind of on the lower end of that um so you can ultimately create a you know what what people who are traditionally look at this stuff as known as an income statement um where let me just walk through some of the key metrics here the 2.

4 billion number is is what I previously showed on the on the screen the cogs is in some ways what what ends up as being interesting this is the cost for a blockchain to actually produce that block space for the consumer and ethereum ultimately has to pay an end user or or or service provider to provide that bandwidth storage compute um and memory to uh to the network to run the service right and so ethereum ultimately pays out uh 1.75 billion of the 2.4 billion uh that they that they earn out to the service providers which is both the token holders as well as um the validators themselves so this first line here is uh the EIP 1559 tip that goes directly to uh the validators and then the rest is um a validator compensation as as well so um that is and if you the other thing if you can look at the gross margin so this is a very a fairly low gross margin business in aggregate which is a a metric that you know obviously investors pay a lot about pay a lot of attention to um September and October you could see those are actually negative gross margin months where because the fee Revenue in September and October on ethereum was low and they were paying out a recurring uh sort of block reward uh they they ultimately had negative gross margins there but in aggregate ethereum kind of averages around 30% uh gross margins on a monthly basis um which is you know comparatively low if you contrast that with arbitrum and you know with other l2s you know arbitr made 63 million in uh in 2023 in transaction fee Revenue they paid out 45 million or 46 million in call data costs down to ethereum so ethereum this 45 million goes into into ethereum's Topline transaction fee Revenue number right and that that averages around again 25 30% gross margin um that is something you know that that that we'll talk about here in a second so what are these systems actually doing um and how much data are they processing and what are what are consumer or what what are the you know L2 consumers paying for L1 block space so they're paying for ethereum Block space somewhere in between ,000 to $2,000 a megabyte um to access ethereum block space and in high you know in in in high months arbitrum is processing about 2.2 kilobytes per second onto uh base base layer L1 and in kind of a stabilized time period may may of 2023 was a comparatively High uh high high month in terms of transaction fee usage and then in August it's kind of more a normalized number at uh 1.2 that's kind of where where it where it hangs out about one 1 kilobyte per second a eum as a system today processes about 14 or 15 kilobytes per second in data throughput um after EIP 4 4844 uh you're going to add another 20 to 25 kilobytes per second in Blob space which will be priced actually differently than than base L uh L1 and so you're going to be expanding that capacity and then obviously the long-term Target for ethereum is uh about 83 kilobytes per second so you could think of that as having basically capacity for 83 arbitrum um um at at kind of fully fully baked um and then pricing congestion based on you know users demand for that so just to highlight and then I'm going to go into a little bit on meev because I think that that stuff is uh quite important to the the study of block space as well but if you look at just how Celestia has performed um in terms of transaction fee uh costs and and so on uh Celestia transactions are running about to5 a megabyte so you could you could say even from the low end of ethereum call data it's about 20x uh cheaper than ethereum call data so if you go back to this arbitrum income statement uh you're looking at this 45 million number divided by 20 and then take the and then basically take that $2.

5 million as cogs on the um on the arbitrum income statement it completely changes the business model of l2s to have cheap data availability that will also be true post 48 uh EIP 4844 uh where they're expanding blob space obviously to 37 um uh 375 kilobytes per block um so to go into the me side a little bit because as I thought about the the um you know the the market for Block space meev is a huge part of that me is a huge consumer of block space and it's a really interesting um Market to analyze as a as a consumer of of this product and this business model so I I I would argue we have two transactions on blockchains the first is uh we have users who exogenously have demand for uh to to complete a transaction that they might want to lend in compound and borrow against a particular asset um that the second type is uh me transactions where there's an endogenous demand from a riskless Arbitrage that a user can extract from the system I believe these are two different classes of transactions that are fundamentally different so one of the qu one of the key questions I want to understand in analyzing block space is how much does type 2 account for my expectation was actually higher than it end up ended up being so in 2023 across the entire meev supply chain um the sort of aggregate dollars earned by the The Searchers the miners from the the priority fee as well as the eth burned as part of that process was 344 million um across the supply chain Searchers made 177% of that and then the network validators plus eth holders made 83 uh% of that validators themselves earned an additional 235 million just from uh me transactions in 2023 so this number was a little bit lower um here's the meev transaction volume on a per week basis paid to the chain so you can see there's two spikes the first is the FTX collapse um in November of 2022 and then the second is the svb crisis this makes sense me transactions are saturating a larger percentage of the overall transaction fee uh you know transactional usage of the uh the ethereum blockchain uh at at that time uh on average we hang around about meev being 10% of the total transaction fees of the system so it's a slightly smaller amount than than uh I ultimately expected um if you break it down by category in terms of the type of me transactions that are happening Liquidations in aggregate in 2023 um were accounted for about uh 7 million of of uh of ethereum revenues Arbitrage accounted for 200 or 125 million of of the sort of me supply chain sandwich are uh sandwich transactions accounted for 210 million and Searcher Builder this is actually there's not great public data on this this is internal Galaxy data and doesn't represent the complete market share of Searcher Builder Arbitrage but this is transactions priced um both based on the binance uh and market price at the time as well as the transactions that actually got executed on chain Searcher Builders which are doing uh centralized exchange decentralized exchange Arbitrage primarily made about a 100 million Topline you can see actually the costs that each one of those players paid out so if you look um Arbitrage transactions are comparatively lower margin than for example liquidation transactions or Searcher Builder transactions so gross margin on liquidations are about 60% uh along with Searcher Builder Dex sex Arbitrage and then on arbitr just normal Atomic onchain Arbitrage the margins for for Searchers is about 18 and 14% um so you can see where what what it would signify is um liquidations or sandwiching and Arbitrage are more compe itive on Chain versus um Searcher Builder kind of Arbitrage opportunities as well as liquidations um and then there's the frequency so if you think about transaction fees as a sort of stream over time liquidations are contributing Less on a on a sort of recurring per week basis but when they happen it's hugely spiky which makes a ton of sense when there's a lot of volatility and then Arbitrage is kind of a a consistent feature of the market um one of the other analysis analyses that that's really interesting and which you know I think factors into our analysis of block spaces block space is not some aggregate thing where you on every day you have a uniform value of a of a block at you know at any particular time the actual value of a position in a block is very different um from the start of a block to the end of a block so this is the aggregate fees paid in 2023 between at every position in a block so the first position of the block the second position of the block the third position of the block all the way down to 100 obviously it tails off five there's not always 500 transactions in a block it's actually pretty uncommon kind of tends to tail off around 100 200 um but you can see 20% of all the revenue made by ethereum or all the transaction fees paid into ethereum happened in the just the first 10 blocks you can see the sort of extremeness of that Spike um at blocks 1 through 10 and and then you kind of tail off I actually didn't understand and I don't so if anybody from the audience has any ideas or thesis of why this is happening I don't I don't understand the dip between transactions about 20 through through 50 um you could see the dip the kind of bodal distribution here uh it surprised me I'm I'm I'm confident the data is correct but if anybody has any ideas it could have to do with uh sandwich um sandwiching Arbitrage happening kind of at at uniform positions in the block um and then the question the final question we want to ask is like what's the contribution of me to fees on a go forward basis right we we identified me is an endogenous transaction type it's created by the usage of or by exogenous demand for actual transactions happening on the Chain how does it you know how do we expect it to to participate in in or contribute to eth transaction fees or blockchain transaction fees on a go forward basis um in general it does about you know 10 to Max 30% of the the sort of overall transaction fees on a chain it's highly correlated as you'd expect with um the total transaction fee volume so on the on the top right you have the weekly ethereum transaction fees paid on a uh uh you know on a particular week and then on the y- axis you have the fees the the fees generated specifically by the meev transactions you can see obviously those are super tightly coupled uh and then the the lower chart is the weekly you know trading volumes uh across all of you know centralized exchange markets versus me fees I would expected them to be slightly more correlated so um these are some of the things you know we're paying close attention to just to wrap up um and to talk a little zoom out and talk a little bit more about block space as a business model so what are what are what are the properties here a block space I believe it is a fundamentally interesting and novel business model that's different from the sale of compute like Amazon ec2 it's the sale of compute with strong Network effects around Capital liquidity application development and and and more um it's prone to network effects there's High net income margins so the actual cost to serve uh serve the block space are all embedded in the cogs the cost of goods sold which are direct cost of the transaction um they are there's no indirect cost of the transaction there's no operational expense overhead to serve that everything it's it's as if Uber was sort of run by by a a network of of drivers and they were they were also running the software to to the network and serving um us that way there's no operating company that's sitting behind it so there's a high propensity for profits here once especially with rollups once you scale gross margins and you cut cut down the cost of call data um and then I think we're just going to benefit from increasing scale over time we're seeing arbitrum and optimism grow 100% year-over-year you know we've seen Avalanche grow 100% year-over-year this is a business model that's doing billions in dollar bill billions of dollars of you know gross you know sort of gross revenue gross merchandise merchandise value and I think it's only growing with the demand for Block space and the demand to use these systems so that's a highle overview of you know how Galax you know at Galaxy we analyze the value of block space and the business model that's really served by these um infrastructure uh layers so that's it thanks everybody