The Saylor Series | Episode 16 | Bitcoin Economics and Evolution
Michael Saylor 0:10
We just finished discussing the proof of work mining network and the seven layers of security: energy, technology, politics, finance, the network itself, the spatial security, the temporal security. So what you have is you have Bitcoin as a decentralized crypto-asset network and the thing that pops up a lot is the question of, What is energy usage look like over time? And is energy usage going to keep increasing as the price of Bitcoin increases? And I’ve seen commentary on this, I think a lot of people get it wrong. They seem to think that as the price of Bitcoin increases the energy usage will increase linearly. If there’s a 100x increase in Bitcoin price there’s a 100x in energy usage. And I think it’s just worthwhile to make the observation that over the past 10 years the mining network has gone from being energy-intensive to being technology-intensive. Another way to say it is that in every single industry you go from being labor-intensive to capital-intensive. When we started with a million people sowing or farming, farming it was a labor-intensive activity. As the capital equipment gets better, first you have horses and carts and ox-carts, and then you have tractors, then you have mega-tractors, then you have factories. All of a sudden, the amount of labor matters a lot less, and the amount of capital equipment matters a lot more. There was a time when 90% of the country was farmers and now just 1–2% of the people in the country produce all the food, because it’s become technology-intensive. The Bitcoin network is similar except substitute for labor, energy. And substitute for capital, technology. If we go back 10 years, it probably took 100x as much energy to generate an exahash as it does right now. An S19 takes 30 megawatts per exahash, but an S9 takes 150 megawatts per exahash, so you’ve got 5x improvement in energy efficiency over one generation of equipment. We’re like on the 7th generation of mining equipment. If you go back 2 or 3 generations then you get that 100x increase in energy intensity. If you look at where we are today and you go forward 10 years, it’s reasonable to expect that you will get improvements in efficiency from three dynamics: you’ve got halvings in the protocol and we have one every 4 years so that means over the course of a decade you have a 5x increase in efficiency from the halvings, then if you have a 4x increase in energy efficiency and you get it twice — or a 5x then a 4x — you get a 20x increase over 2 more generations of hardware, and now you’ve got a 20x times a 5x or about a 100x increase in efficiency. So Bitcoin price could go up by 100x and the network efficiency would go up by 100x and the energy consumption could be flat. And that’s an important thing to keep in mind because people sometimes think, Well energy is used to secure the network — No it’s crypto-energy. It’s encrypted energy that’s used to secure the network. And in order to get it from raw energy to crypto-energy you have to run it through a SHA-256 miner that’s properly engineered in a heat-sink. And so as the heat engineering improves, the miners get more efficient, as the semi-conductor engineering for ASICs improves, the network gets more efficient. And what you have in that dynamic is a never-struggle between brute-force and technique. It kind of reminds you of something — like nature, and competition. There’s two ways to do things, right? You either use raw labor, or you use technology. And it’s important to have that dynamic, or that yin and yang, because when the technologists get lazy and they stop improving, then the raw material or the brute-force overwhelms and the control of the network shifts back to the energy holders. But to the extent that the technologists upgrade their engineering facilities, their heat technology, their chip technology, then the energy becomes less important. The technology becomes more important. And the combination of both of these things require free-flowing capital. In a free market the capital is continually seeking the best use. Should I invest money in creating the next generation of SHA-256 miners? Should I invest money in engineering a more efficient Bitcoin mine with immersion cooling? Should I invest money in commercializing more energy? Or plugging in more energy and just manufacturing the same design over and over again? So what you have is this nice delicate dance or balance of power. But a model you can think of is when we first started Bitcoin mining, it was all energy-intensive. We were using off-the-shelf computer equipment and you were throwing raw and power and raw commodity materials at it. It’s rotated through a set of generations of hashing equipment. So now you need energy but the limiting factor is not with the energy, the limiting factor on generating hashes is the mining rigs. And maybe you could say the limiting factors on doing this well are the mining rigs properly installed in the right mining center with the right cooling technology. And if you look forward another decade, you’ll see it’s going to be much more technical-intensive. You could have all the energy in the world, but you’re not going to be able to generate crypto-hashes. The break-even point of an S19 is like 40 cents a kilowatt-hour. You could pay  cents a kilowatt-hour for electricity and make money. The break-even point for an S9 is 8 or 9 cents a kilowatt-hour. If you have 20 cent power you can’t make money—you’ve got to turn it off. The break-even point for the generation of equipment before that is like 2 cents per kilowatt-hour. You can’t pay 3, you have to turn it off. And the generation before that you’re down to half a penny a kilowatt-hour. So if you’re looking to run antiquated equipment to generate hashes you have to be stealing the power. If you have free energy and it’s literally stolen or it’s given away, you can run 8 year old or 5 year old equipment. But if you’re paying your way, you have to run modern equipment. So this is a dynamic model but it’s important because it’s a dynamic evolution of the sophistication of the Bitcoin security network that has Darwinian overtones. Because the Bitcoin miner that just buys a bunch of equipment 5 years ago and stops upgrading becomes obsolete. You won’t last 10 years with that equipment. I mean your break-even point becomes a tenth of a penny once you’re four generations behind. So this is like nature healing itself. It’s a very natural process. If you’re out of touch and you’re 1,000 miles away, the free market somewhere else is going to keep moving the state of the art forward, and should you isolate yourself you will find that you can’t generate enough hashpower to participate in the revenues and the transaction fees. And it’s the network’s way of simply squeezing you off the network. Now what means is Bitcoin miners are organic creatures with a negative feedback loop. In essence a market mechanism. The difficulty adjustment is not just tactically every 2 weeks in the protocol. The difficulty adjustment is in the market, because if you don’t upgrade your hardware every 4 years you become uncompetitive. And you can’t upgrade your hardware every 4 years unless you were a responsible custodian of your capital. If you spent all your money, if you didn’t save any money, if you’re not trustworthy, if you don’t have credit, you can’t buy any equipment. And so to stay competitive on the Bitcoin network you have to be credit-worthy, you have to be competent, and you have to be credible. Someone that has the technology has to be willing to sell to you. And so I would say that the dynamic nature of the network is: you have this competition between miner, operations, and the technology providers, and the energy sources, and the political jurisdictions, and you have to be upgrading, and you probably need to be improving your efficiency by a factor of 1.5–2x every year. You’re subject to Moore’s Law and this competitive Darwinian pressure. The result of that is: energy consumption is going to fall per exahash. We’ll go from 150 megawatts an exahash to 30 megawatts an exahash, to 5 megawatts an exahash, to 1 megawatt an exahash, to half a megawatt — and on down. We’re just going to move down this efficiency curve. You could imagine if I gave you this supercomputer the size of a sugar cube — look at the amount of power on a modern semi-conductor, right? Think about what’s in the latest iPhone chip. We just keep compressing computing power and we find a way. And human ingenuity is like that. They will keep finding a way to create more hashpower using better technique. So that suggests that energy consumption will probably increase non-linearly, like with the log of the price. If the price goes up by a factor of 10x, energy consumption might go up by a factor of 2, 2.5, 3x. At some point you start to think you will roll over — you’ll peak. We might have already done it. You peak energy consumption and then you taper off or it holds constant as the hashrate increases, and instead of a Bitcoin miner spending half of their budget on energy, you roll over to spending half of your budget on capital. You’re buying capital and amortizing it, and pretty soon your variable cost on energy is 5% and 40% is capital equipment, and what you’ve got is not an energy war to see who controls the network, you’ve got a technology to see who controls the network. Why is that good? Well because energy is a raw material in the universe, but 7th generation SHA-256 miners are specialty equipment. Everybody on Earth can find some energy and throw it at the problem, but throwing the 11th generation of SHA-256 mining equipment at the problem is something that probably no one’s going to be able to do unless they spent a decade or two decades engineering that equipment and thinking about it. So you have a specialization in the same way that John Deere tractors are specialized after 100 years. And if you went back to a farmer in 1850 and you describe what a farmer in 2020 can do, it’s like night and day. That’s what’s going on with the network. And I think it’s self-healing, self-sealing, self-correcting, and the combination of the Darwinian and Adam Smith capitalist competition is a critical advantage for Bitcoin versus a proof of stake network. Because in the proof of stake network, you pretty much turn off energy competition, you turn off semi-conducter competition and innovation, you turn off engineering and heat engineering innovation, you turn off capital financial competition, you turn off the political element, and that means that I can get fat, dumb, and happy. I can just post a billion worth of my tokens and go to sleep for a decade and no one is trying to make a network so it’s not being tested, there’s no stressing of it, and ultimately the problem with that is you’re going to suffer in integrity and durability.
Robert Breedlove 15:15
So we’re again at this point where the Bitcoin mining network itself is a free market in and unto itself and it reflects many of these properties of capitalism. So there’s this yin and yang of efficiency and magnitude going back and forth. And to your point that’s reflected in many markets historically where we shift from labor intensity to capital intensity. And this basically means as the capital becomes more plentiful and effective at amplifying labor, you need less labor to accomplish the same results. This is a natural, capitalistic progression. Let me ask you this: the useful life of a miner I’m not sure what that is —
4 to 5 years.
Robert Breedlove 16:02
So we have the older generations rolling to cheaper energy to remain in the marketplace.
I can steal energy. If I can bootleg energy I can run older generation miners profitably because my variable cost is zero.
Robert Breedlove 16:15
And then so newer generations would then be allocated to more expensive energy resources initially. Is this an aspect of the gradually increasing fully amortized cost to produce each Bitcoin? Because there’s more cap-ex — I mean maybe the cap-ex, op-ex mix is changing, but the overall cost to produce each Bitcoin is rising, which is putting upward pressure on its price in the marketplace. Is that another way to look at this?
Michael Saylor 16:48
That’s not clear to me. What’s the price of energy? Is it going up or going down? Maybe the price of energy is going up and then it’s going down. What’s the price of semi-conductors? Is it going up or is it going down? Well sometimes it’s going up. It’s going up if there’s a monopoly on the semi-conductor chips, and then it’s going down when someone else enters the market. An Intel 386 chip is pretty cheap now, right? So chip prices are coming down but the next generation is going up. But again it’s a competitive thing. What’s the price of semi-conductors? If there’s only one provider and it’s a monopoly then you could say, Well the price of your semi-conductors is going to go up every generation forever and the cost of Bitcoin mining is going to go up. But there isn’t [only] one semi-conductor company. There’s multiple. If there is one, we’re back to Jeff Bezos’s statement, Your margin is my opportunity. Bitcoin is an open market for mining, so anybody can engineer a SHA-256 chip. There is no one type of Bitcoin miner. There are people slapping these things in the back of trucks and driving them around and plugging them into bootlegged power lines, right? It’s not very efficient. There are people that slap 500 Bitcoin miners into a container and they just drop it on top of a pad. Does it work? Sure it does. Is it properly heat-engineered? No. Does it get hot? Very. What’s the consequences of running hot? You have to turn off the mining equipment some of the time so you lose efficiency, and they burn out and their useful life is not 4 years, it’s 3 years. Or it’s 2 years. So there’s a competition here and just like there’s competition for energy, will energy go up forever? No energy could go to zero. If I invent cold fusion. Will semi-conductors go up forever? No. At some point Bitmain raises their prices and then everybody starts talking to Intel. You go to another semi-conductor company and you say, Well look Bitmain’s tripled their prices and now you can make this much. And now you draw somebody else into the space. So I think there’s a dynamic equilibrium between the engineering and the semi-conductor and the energy. And the best capitalized companies, they can buy the new equipment. Like the publicly traded North American companies, they will go and buy all the new shiny equipment, and poorly capitalized companies will run the old equipment. And then you will roll forward like that. It’s not clear to me how much it will cost to create a Bitcoin, because that’s a function of — maybe the profitability of Bitcoin is a function of the rate at which we add hashpower versus the rate at which we add Bitcoin HODLers. If we increase the number of Bitcoin HODLers by a factor of 10x, and we increase the hashpower by a factor of 2x, then the profitability — the revenue per exahash — is going to go up by a factor of 5x, right? And if we increase the exahash by a factor of 10x and the HODLers by a factor of 2x, the profitability is going to deteriorate. So the profitability of the Bitcoin mining is going to have an impact on the rate of development of Bitcoin mining semi-conductor ASICs, right?
Robert Breedlove 20:52
So the total op-ex and cap-ex going into the Bitcoin network at a halving — that total allocation of expense goes from producing call it 900 Bitcoin a day —
The revenue gets cut not quite in half.
Robert Breedlove 21:10
So the production cost per Bitcoin is effectively doubled. And then my thinking was that this higher cost to produce each Bitcoin is actually incentivizing miners to HODL. They don’t want to sell below cost of production. So that’s what’s bootstrapping Bitcoin’s price upward.
Michael Saylor 21:29
I can see once every 4 years the energy cost doubles. But if once every 4 years I upgrade the equipment and it’s 5x faster, now the question is: how much did you pay for the equipment? And the answer is: if you paid a lot for the equipment — now amortize the cost of the equipment into the cost of the Bitcoin, and you get an answer. But that’s a price that a vendor charges you, which could be triple or half depending upon the competition. In a very competitive market, well let’s say it a different way — if you’re buying Intel grade 386–46 chips and slapping them into your home appliance, what is the price of the intelligence that you put inside your toaster? As a percentage of the cost of the toaster? It’s pretty cheap, right? Like at some point the cost of intelligence starts to drop because it’s a commodity market, AMD drove down the prices and so everybody can put intelligence into their appliance because of competition. If there was no competition it might go the opposite way. So I think that protocol guarantees us through halvings that we’re going to have to get 5x as efficient every decade, but the other dynamics like the rate of technology advance go even faster. And they’re just as much. I do agree with you on this one idea: we’re in a gold rush right now. Or a Bitcoin rush. Between now and like 2035 — in 2035 we’ll have mined 99% of the Bitcoin. So you’ve got 14 years, and during the 14 years the block rewards are pretty high relatively speaking. And so it’s very lucrative to be in the business, but the business is going to become less profitable, likely. At the very least we know it’s going to rotate into a transaction-fee business. So instead of 90% block reward 10% transaction fee, you could expect 90% transaction fee 10% block reward. And the transaction fees are also going to scale more likely with the log of the price, not linear to the price. If I want to move a million dollars I might pay you $10, and if I want to move $10 million I would pay $12 but I’m not going to pay $100. I’m just going to overbid the million dollar buyer. And if I want to move $50 million I’ll bid $14. So the transaction fees will increase with the number of transactions on the base layer, and that’ll be a dynamic. But the block rewards go away. And so if you’re a Bitcoin miner it’s a very lucrative high-growth business right now, but it will move toward an efficient transaction security network later, and you would do well to buy Bitcoin. That’s why it makes sense to HODL Bitcoin if you’re a Bitcoin miner, because it’s your hedge against two things: it’s a hedge against the inevitable protocol — which the protocol says Bitcoin mining is going to get 90% less lucrative over the course, cut in half once, cut in half twice. Over the course of 12 years you get cut in half three times. So it’s a hedge against that but that’s the long term. It’s also a hedge against the near-term hashrate explosion. If you’re mining Bitcoin right now and if somebody comes out of the blue and they bring a lot of hashrate on the network faster than you expected then your market share gets cut. In that situation [HODLing] Bitcoin may win. If everybody believes that Bitcoin will win and there’s a flood of people that enter the mining business, you want to own Bitcoin. Mining’s going to get competitive, Bitcoin’s going to get more valuable. So you definitely don’t want to be in a situation where you’ve mined Bitcoin, you sell all your Bitcoin for cash — in that case you’re attempting to run a cash business in 16 years when the revenue’s been cut in half four times. Presumably if you cut a 100 to 50 to 25 to 12 to 6, when you’re generating 6% of the revenue and everybody else has had 16 years to get in the game, it’s going to be more of a commodity business. Now that’s not to say that a Bitcoin miner can’t compete and evolve, it may be a great business then too. It may be that Bitcoin miners evolve to be running Lightning nodes or running layer 2 platforms of other sorts. It might be that as a Bitcoin miner you have a huge Bitcoin treasury and then you can generate yield on it. It might be that there are other applications. We’ll talk about some. It’s very logical for Bitcoin miners maybe to get into some of these other application areas that will pop up in time. And there are technology possibilities here. So I would say the illogical thing to do in business is you don’t worry about what happened 16–20 years out if you can make a ton of money now. I mean you might very well make enough money now to have 10, 20, 30, 100 billion dollars of capital then, at which point you can go buy something else to get you in the business. And my other point I guess here is: yes it’s true, Bitcoin miners should keep Bitcoin. But it’s also true that non-Bitcoin miners should keep Bitcoin. Like if you believe in Bitcoin, then the logic follows that no matter business you’re in, if you have operating income, you should invest in Bitcoin. And if you are able to generate financing, if you can raise debt or equity then you should also raise debt and invest that in Bitcoin. That would be the case for other companies. Microstrategy looks at it that way and we’re not even a Bitcoin miner. But that’s my thought on the Bitcoin mining network and the dynamic model over time. I think that some people intentionally misunderstand this, like proof of stakers, they intentionally misunderstand because if they misunderstand it then they can assume that energy consumption is linear, and then they can play the ESG card and pretend that they’re virtue-signaling and say that they’re doing something good for the environment because they’re not using energy. But the truth is that they’re not using energy nor are they using technology. They’re not using hardware technology, engineering technology, or energy, nor are they submitting their network to competition, and they’re not using external capital, political or financial. What they’re doing is they’re creating a virtual protocol in order to create security. And even there, if you’re going to rely upon a virtual protocol to create security you’d be better off to stake the protocol with an asset that is outside of the network, that is external to the network. So for example, the Lightning Network makes a lot more sense because it’s staked with Bitcoin, than if it was staked with “Lightning Coin.”
Robert Breedlove 29:57
Yeah and it’s paying the revenues for services rendered, which is actually the routing transactions, versus just how much Bitcoin you hold — that’s not how it works. So I think the key point here is that the proof of work energy expenditure is actually transforming what would just be a video game asset into a macroeconomic asset.
Michael Saylor 30:22
It’s creating a real-world asset. It’s your connection to thermodynamic reality and socio-political reality. You can’t be in denial of those two. What people think of you matters. If the person is the police officer on the beat or runs the country, what they think of you matters. And what nature thinks of you matters. If you decide to step off a cliff by looking the wrong way, nature’s opinion matters whether or not you respect nature or not. And so it’s very important that if you’re building something that you want to last for 1,000 years, you respect politics, and you respect thermodynamics, and you respect physics. And proof of work is this very creative invention to continually connect and synchronize the Bitcoin network with political, physical reality. Never-ending evolution. Every 4 years, every 2 years. A new miner, a new chip, a new place. Throw away the old, the old generation has to die so the new generation can form, so that the creature can evolve. Otherwise the creature stagnates and becomes progressively more fragile until it’s no longer capable of competing in the real world.
Robert Breedlove 32:04
And that energetic anchor to these realities — energetic which begets socio-political reality — this is what’s keeping all network participants honest and accountable. There’s this synchronization and rule set imposed upon them that they can’t avoid. It’s like gravity. You just can’t avoid it.
Michael Saylor 32:28
Yeah it’s quite a wonderful thing. It’s a crypto-universe. Satoshi played God, you either call it creating your own universe and setting the planets in motion by defining the space-time constant, or the other metaphor is you released a creature into cyberspace and set the genetic DNA of the creature. You controlled how it will procreate. And once you release the thing, then it spreads as some kind of swarm life-form continually evolving. So I think with that we conclude that the Bitcoin mining network and the Bitcoin network in general is quite a wonderful thing. Now the next question is: how does it scale? If we go beyond the network, the primary purpose of the Bitcoin mining network is to provide security and enforce protocol integrity and durability of the system. And it does that very well. But so how do we actually scale out to provide hundreds of billions of transactions a day to billions of people on the planet at the speed of light using the latest computer technology? That’s I think where a lot of people fall down, they just want to look at the base chain and say, Well it moves 7 transactions a seconds and so you can’t scale. But they lack the imagination to understand the consequences of the layering. So the Bitcoin monetary network, it scales based upon platforms at the layer 2 level, and then applications above that level, and you could call them layer 3 applications or you could call them applications on top of applications on top of applications. So you could have layers 4, 5, 6, and 7 and you could have interplay between the applications, but for the purposes of our discussion let’s just call them layer 2 platforms and layer 3 applications in order to keep from getting too confusing in our semantics. So what’s the most important one? Well Lightning is a very interesting and maybe the most important layer 2 platform. And the idea behind Lightning is: I want something to be exponentially faster and exponentially cheaper and in return I’m willing to secure exponentially less money. So if I have only $100 at risk, then there’s no reason why I couldn’t do 10 million transactions on that channel. And that totally makes sense because as you scale out, if you look at the transactions that 8 billion people need to make on this planet every single day, the majority of the transactions are actually of value less than one dollar. In fact there are plenty of transactions that could be valued in pennies. And then the big transactions might be $10, $100, or $1,000. Probably if you were to look at the transaction stream on the Visa network or the MasterCard network, the average check is like $28 or something. But there’s billions and billions and billions of them. So you don’t really need the final settlement security of Bitcoin, because that’s the highest level of security in the world. And you’re getting that to move a billion dollars. You probably only need to move a billion dollars around occasionally. If I moved seven nuclear-powered aircraft carriers per second anywhere in the universe, I could probably win whatever war I wanted to win if that’s what I’m using my teleportation power for. So Bitcoin is good for teleporting large chunks of value, but for all of the routine work, you want something smaller. So the logical thing to do is — the way Lightning gets at it is you’re putting liquidity in the channel and the channel is at risk, but the overall blockchain isn’t. The beauty is you get to keep your keys and keep custody of that. So that a very fascinating thing. You can build the universe on Lightning. Presumably, you could give 8 billion people Lightning wallet or wallets — more than one — you can build Lightning into PayPal and Square Cash and Apple Pay and Google Pay and Facebook and every messaging app, and you could use it at all scale, not just to move money but to move anything. The other day I did a 1,000 satoshi transaction on Lightning for 1 satoshi in 1 second. So Lightning is an obvious layer 2. And the simple design idea is: I create a channel with a million times less Bitcoin in it, and I move it a million times faster. And that works fine. And there are other approaches of course, but you can give up all of the proof of work because you don’t have as much at stake. And you’re creating a staked network but you’re staking it with an external asset that derives its value from assets external to it — energy capital and political capital and technical capital are flowing into Bitcoin capital and Bitcoin capital is flowing into the Lightning Network in order to secure it. So once you get that idea, you realize that a layer 1, layer 2 solution is a lot better idea than a higher performance layer 1 solution. Because making the layer 1 twice as fast or 3x as fast or 10x-100x as fast doesn’t get you to a million times as fast. It’s no different than the common-sense way that human beings solve every problem: if you had a million dollars in the bank and you were going out on a Saturday and you needed to spend money quickly, you would take $100 in cash and you would break it into 20 $5 bills, you put it in your wallet, and you would rest assured that you can’t lose more than $100 and you would leave the rest of the money locked up in the bank. And the money in the bank takes 48 hours to get at and takes lots of degrees of authentication that’s behind 3-feet of steel, and the money in your wallet is in your front pocket and it takes 1 second to get at and occasionally you drop a $5 bill on the floor when you’re drunk and life goes on. And if somebody said you had to take the entire million dollars with you out to a night club with you every Saturday night you would say, That’s pretty stupid. And if they said, Well you know the bank vault doors they’re too heavy, we’re just going to have to re-engineer them to make them 1,000x lighter so that you can open the doors faster so that you can access the million dollars on Saturday night while you’re drinking — you might think, Well maybe it wasn’t such a good idea to actually have access to all my money while I was out drinking on a Saturday night? And the same is true with the idea of a base layer. Maybe it’s not such a good idea to have that many transactions on the base layer, because every single moving part is just something to break. So you’ve got yourself too many moving parts — you don’t want it. So the layer 2 platform is this idea that what I want to do is I want to just create a cash, if you will, of a much smaller amount of value that is much less at risk that I can afford to move much quicker and I don’t need the same degree of security on it. And Lightning is not the only layer 2 platform you can conceive of. Another crypto-network could be a layer 2 platform. In theory you could spin up a proof of stake and you could stake it with Bitcoin and it wouldn’t be a theoretically different thing. You can also spin up a crypto-network proof of stake and then you can move Bitcoin through it, but if you’re using the native token and moving Bitcoin then your Bitcoin is only as secure as the native token. And of course if the token is ginned out of thin air — Yoyo-coin — then that’s a problem. Ultimately the real issue with layer 2s is you’re moving a portion of value into that layer 2, trusting the layer 2 in order to get performance or functionality. So you can do it, you can do it with Lightning, and the logical thing you do is you stake with Bitcoin and then you reduce the risk to the channel liquidity. That’s a way to do it and that’s logical. Another way to do it is to build some kind of system like maybe if I was an exchange and I had Bitcoin on the exchange and I created an API to the exchange that was programmable. Then in a way the exchange is holding the Bitcoin, the API is providing you with a very fast speed. Now you’ve taken a different risk. You’ve taken exchange risk, which is some counterparty risk of some sort. If you use a Paxos or a NYDIG platform those are more conventional platforms to build applications on top of. Like NYDIG has a platform for you to build a credit card plugged into Bitcoin. And Paxos has a platform plugged into PayPal to their application. So if you want that kind of platform you plug the mobile application into one of those platforms — there is some risk with that counterparty — it is limited to the amount of Bitcoin you’re handling on the platform. And you can do all sorts of things to mitigate the risk, but clearly there is a need for layer 2 platforms. And there isn’t a need for just one layer 2 platform. Like Lightning is a compelling, decentralized layer 2 platform because it’s decentralized. But there are a lot of layer 2 platforms that will be centralized because they need to be regulated in order to meet with regulatory compliance obligations. And there are a lot of counterparties, like a credit card company might prefer to work with a centralized layer 2 than a decentralized layer 2 because they have regulatory constraints.
Robert Breedlove 46:01
And this is similarly rooted in something very fundamental, as proof of work, because there’s this fundamental trade-off between security and freedom, typically. And so in money we’re saying that the security model of Bitcoin is its decentralization, but with that comes a lot of work. It doesn’t do many transactions per second. But what you can do is abstract that Bitcoin to a more centralized database whether it’s a custodian or a Lightning — a Lightning is kind of a decentralized centralizing force — you pick up all this functionality but you’re giving up some of the trust-minimization you get at the base layer.
Michael Saylor 46:42
I would say that if your focus is on property, and what you want is digital property, you want the system optimized for durability and integrity over time. And performance and functionality and compliance are not on your list. Your list is durability, integrity, immortality. Very simple. But as you move towards applications and you move away from property towards applications, you either have to optimize for functionality — like if there is no functionality there is no application — or you have to optimize for performance. If I can’t pay for the coffee within 1 second I can’t pay for the coffee. Or you have to optimize for compliance. If I really wanted to issue an insurance policy or I want to issue a security or if I want to issue a yield token, and it’s illegal in a certain state, then if I want to do it I have to actually comply. And compliance pops up with stablecoins. Compliance pops up with DeFi exchanges. Compliance pops up with derivatives. Compliance with pop up with anything that looks like a security token. These are all applications, but is there a future for those applications on top of Bitcoin? Yeah. But Bitcoin is not the application. The application will be built on a layer 2 platform. Or it’ll just be built naked against the Bitcoin. Maybe you don’t use a platform, like I can implement my Bitcoin credit card using the NYDIG platform, or I could hire an army of programmers and I could build it one-off, right? There’s this little battle of: Do I build a custom app or do I build an app using your SDK? So platforms are going to be operating system-type things with an SDK, and it might be that it looks like an AWS where they spin up services. You could imagine an AWS spinning up a whole package — like a Lightning node — if I could spin up my Lightning nodes and run them maybe I would do that. Maybe I wouldn’t do that. Probably Lightning is not the best example because people are looking for something totally decentralized there, but a better example would just be an SDK that allows someone to deploy a mobile app that has Lightning and Bitcoin money transfers embedded in the mobile app and I just want to do it quick and easy.
Robert Breedlove 49:50
There’s a continuum where we have this totally decentralized layer 2 in Lightning, maybe we have SDKs in the middle versus a fully centralized solution via NYDIG or someone else. But it just speaks to the versatility of Bitcoin which again you can’t get from a gold. You can’t get this versatility of application layer with something like gold.
Michael Saylor 50:12
And there’s going to be competition, right? Massive competition with regard to: what are the SDKs in the layer 2 toolkits? And even Lightning has Lightning Labs developing a toolkit to help you with Lightning. So that’s competition, and it’s interesting but it gets theoretical in a way so it’s actually probably more instructive to move down to the applications themselves and talk about: what are the applications of Bitcoin that scale the system? Because these are the things that make the different. And once you start to think about the range of applications then you figure out what you might build into your platform if you’re trying to create a business hosting those applications. Let’s talk about that. The obvious application is the individual HODLer that owns Bitcoin or just HODLs it in cold storage for long periods of time. It’s like the family or the individual. I think that’s pretty well understood and that’s been pioneered for a while. The thing that’s varying is the way that that individual chooses to HODL the Bitcoin. And that’s where it gets really interesting. For example, one thing that I want to start with is this observation that one interesting application of Bitcoin is a bond. You can actually create a derivative or a security of Bitcoin and scale the network with the bond. So if a government owns Bitcoin, the government is the customer. And if the government then buys the Bitcoin and then issues sovereign debt, the sovereign debt becomes a Bitcoin derivative. So you’re creating credit from Bitcoin. Bitcoin is the base layer money, and the debt or the credit or the bond of the government is the layer 2 money. It could be backed in whole by Bitcoin like 100% or it could be backed in part. If a municipality like a city buys Bitcoin and issues municipal bonds that becomes another form of an application of Bitcoin. If an agency like Fanny Mae or Freddie Mac or any international agency — United Nations or the like — if they were to buy Bitcoin and then issue any kind of bonds, it’s another example of a derivative. When a corporation buys Bitcoin like Microstrategy and then we issue a bond that’s backed by the Bitcoin, we created a derivative. And you could even take Bitcoin — there’s $700 billion of this Bitcoin out there — and you could issue asset bonds. Call it Bitcoin-backed securities. They used to be called mortgage-backed securities. And with mortgage-backed securities we’d take a bunch of heterogeneous assets and we’d securitize them into a note, and so imagine you had 21 million identical houses called Bitcoins and you decided that you were going to create a bond backed by 11 or 37 of them — an asset-backed bond. All of those are applications of Bitcoin. They’re not applications the way a computer scientist thinks of them. They’re financial applications of Bitcoin. You can have technical applications of Bitcoin that are running on mobile devices like mobile apps, you could have web applications, you could have financial applications of Bitcoin. And a lot of times when people think about scaling, they only think about the technical applications and they don’t really think about the financial applications. If we move on the next layer or next type of application, think about mobile payments. A mobile app that does payments is something you can plug into Bitcoin, like Square and PayPal have plugged mobile apps into Bitcoin. What does the app do? Maybe it lets you buy Bicoin, sell Bitcoin, send Bitcoin, send Bitcoin via the Lightning Network like the Muun wallet — that’s an application — but maybe it lets you send Bitcoin on its own proprietary network like Square Cash Tags. You can send Bitcoin between one Cash Tag and another Cash Tag instantly for free. You could send Bitcoin between any two mobile apps within the network using the handle of the network. If you were doing that, then probably the application provider custodies some of the Bitcoin and they’re moving it and they become a fractional bank of Bitcoin. Bitcoin becomes the central bank in cyberspace. Our friend Ross Stevens would say the Decentral Bank in cyberspace. All of these other mobile apps or websites become fractional banks plugged into the central bank of cyberspace. And they custody some amount of Bitcoin. The mobile payments space is particularly promising, because all these mobile applications are one step removed from being mobile banks, and there’s no reason why Facebook and Apple and Google and the like don’t eventually become mobile banks. It was a matter of time before they let you send photos, and then at some point they decided to let you send videos. And then they decided to give you emojis. And then they decided they’d let you send audio files. And in a way you could think of Bitcoin as just another file type.
Yeah it’s all media.
Michael Saylor 56:50
So it’s kind of inevitable. Now you start with mobile payments and then when you start to think about moving that around and the services that consumers want, you’re now into retail banking. The retail banking applications are in essence savings accounts and credit lines based on Bitcoin for consumers. You’ve got billions of people that presumably want a wallet with an asset in it, and the asset would be Bitcoin. Then they want to borrow against it. So you want to be able to draw down a credit line at an interest rate. So I carry around $10,000 in Bitcoin, I borrow $1,000, I pay 4% interest. Now I’ve drawn the credit line in the local fiat currency in question — Euros, Dollars, Yen — I pay the interest whatever that is, and then I spend the money. That’s pretty popular. I mean credit cards are quite popular, right? How many credit cards are there in the world? So we’re really talking about credit cards drawn against Bitcoin property instead of unsecured credit lines. The opposite is also the case: maybe the consumers want savings accounts that generate yield. So I have Bitcoin, I either HODL the Bitcoin, or I move the Bitcoin into a yield-generating account, or I move the Bitcoin into a collateral line and I use it and pledge it as credit. Now is that the same everywhere in the world? No not really. You’re going to have different regulators in every jurisdiction. Every state they’re going to tell you whether you need a money transfer license or a banking license or whether or not you need a securities registration in order to give yield or give loans or move money around. Of course, there are always nuances. I think this is one of the nuances that caused El Salvador to designate Bitcoin as legal tender so that they could easily move Bitcoin around in mobile apps. In a place where the government has collapsed, then the solution is going to be a mobile wallet with Lightning because there isn’t a regulator. And wherever you have the kind of weak governance then it’s not going to matter. As the governance gets stronger then there are going to be questions of: what’s your compliance requirements to do a retail banking application? It looks like it’s literally different state to state, country to country, jurisdiction to jurisdiction and evolving right now. That’s one of the reasons why maybe centralized applications — CeFi — will actually beat the DeFi, because the message of DeFi is: Oh it’s really expensive going through all this KYC. Yes that’s true, but it’s also illegal not to. So the real question will be, Can you do it and how much validation do you have to go through to do it, and how compliant do you want to be.
Robert Breedlove 1:00:20
Hey everybody! So that was Episode 16.