Epicenter - Learn about Crypto, Blockchain, Ethereum, Bitcoin and Distributed Technologies - Dave Carlson, Jez San, Marc Aafjes, Naveed Sherwani & Timo Hanke: CoinSummit London – Pannel: The Ever Changing Landscape of Bitcoin Mining

Episode Date: August 28, 2014

Today’s panel discussion is called “The ever changing landscape of Bitcoin mining” where moderator Jez San is joined by some of the most prominent figures in the mining industry, Marc Aafjes of Bitfury, Dave Carlson of Megabigpower, Timo Hanke of Cointerra and Naveed Sherwani of Peernova. The discussion revolves around several key topics, notably, everyone seems to agree the industry should move away from the term mining and start calling it transaction processing. They also discuss the rapid evolution of ASIC processors in the last year, the importance of improving the overall power efficiency of the network and where ASIC production is heading. They also talk about a highly debated subject which is the shift from home mining to centralised data centers. The panels ends with predictions on the total hashing power of the network at the end of this year. This episode is hosted by Brian Fabian Crain and Sébastien Couture. Show notes and listening options: epicenter.tv/036

Welcome to Epicenter Bitcoin, the show which talks about the technologies, projects, and startups driving decentralization and the global cryptocurrency revolution. My name is Sebastian Couture. And I'm Brian Fabian Crane. On July 10th and 11th, we were in London for the Coin Summit Conference. This two-day event gathered approximately 250 investors, entrepreneurs, and developers to discuss some of the most important issues facing the Bitcoin and cryptocurrency ecosystem. Today's panel discussion is called the ever-changing landscape. of Bitcoin mining, where moderator Jess San is joined by some of the most prominent figures in the mining industry, Mark Athias of Bitpuri, Dave Carlson of Megabikpower, Timohanka of

Cointara, and Navit Sharwani of Piranova. The discussion revolves around several key topics, notably everyone seems to agree the mining industry should move away from the term mining and start calling it transaction processing, and you'll hear the reasons why. They also discussed the rapid evolution of ASIC processors in the last year, the importance of improving the overall power efficiency of the network and where ASIC production seems to be heading. They also talk about a highly debated subject, which is the shift from home mining to centralize data centers, and the panel ends with predictions on the total hashing power of the network at the end of this year.

So, hello again, everyone. We have a much bigger panel now. So I'm joined by some of the most successful and hardworking people in the Bitcoin mining world. And we've come up with a few interesting questions, but let's start with, oh, and we're not going to go around, Robin. We're going to be selective. What is mining and why is it important for the Bitcoin ecosystem? Anyone want to take a stab at that one?

I feel pretty strongly about that one. You guys can hear me, okay? Transaction processing is, that's mining. Mining is basically just the processing of transactions into the blockchain. What could be more central to Bitcoin? Everybody who has a little bit of Bitcoin here, a startup business, intends to invest, they depend on a miner somewhere in the world to add these transactions to the blockchain. So I see it as the epicenter of this happening that is Bitcoin.

And that's what caused me to dive into Bitcoin mining two years ago. Yeah. Mining is certainly the key invention of Bitcoin that allows to come to a consensus in a distributed computer network. That's the key invention. But, well, by the way, by the term mining, we mean at least two things. One is related to the blockchain technology, which is the process of creating a new block. And by doing that clearing transactions that have appeared in the time window since the previous block. And in terms of the currency, it also means the process of creating new currency. And Bitcoin so beautifully links the two things together and creates the incentive to actually build the mining network. So miners get revenue from two sources. They get the block reward for creating new blocks, and they also get a transaction fee. So miners are compensated for their computing power.

Actually, I could say I think if we could stop using the word mining and start calling in transaction processing. Yeah, mining is definitely transaction processing. So I think it really would help the investors. It would help the users and it will help the regulators what we're doing because mining gives a very different kind of a mental image as opposed to what we do. I mean, we are not mining. We are trying to process transactions and add legitimate blocks to the blockchain, right?

So that would be my call to the folks would be that let's see if we can invent a better terminology for what we do. So in the past, mining was done using software running on CPUs, and eventually that was overtaken by running the software on graphics cards. And in the last year or so, it's now been overtaken with ASICs. Does anyone want to give a brief history of the ASIC technology? Maybe I can take a shot at it, considering that I have a little bit of a background in ASIC. I think A6 started somewhere in last March time frame where people started designing this in 55 nanometer or 40 nanometer. By 55 nanometer, these are different technology terms that we use in semiconductor business to explain how advanced the process technology is. And so every few years, you know, technology advances.

So in last year, in the beginning of the year, people started developing A6 in 55 nanometer. in a semiconductor business, it takes about two and a half years to go to the next technology node. In Bitcoin world, we went from 55 to 40 to 28 nanometers in less than, I would say, nine months or so. I think some even did 130 and 110 nanometers.

In between, yes. So that is true. So point really is that I think the advance of technology in Bitcoin is much faster than in the semiconductor business. So today we are at 28, going to 20 and 16. And the reason for the transition from CPUs to GPUs to

ASECs was to become more efficient and lower... Yeah, so if you use a CPU, for example, you're using probably 3% or 2.5% of the silicon area for doing what you are doing. When you're doing ASIC, 100% of the area is being used for doing... So, a dedicated ASIC is, you know,

much, much more efficient in doing so. Yeah, and I think, to answer that, I think the result of that is actually that the network has become much more secure as a result of using ASICs. Just to give you a sense already, the beginning of this year, some computations showed that the combined hashing power on the Bitcoin network was the equivalent of the top 500 supercomputers in the world combined. And that is the only result of having those very efficient chips focusing on that.

It creates stability and security for the network open. And the competitive nature of the mining network encourages people to keep adding. This is all brought on by the block rewards. Block reward subsidizes the growth of the network. and it's designed inherently to promote extremely fast growth because the faster the network gets very large, the faster it's fully secured. So one of the controversial topics in the altcoin world is trying to design proof of work algorithms that are ASIC proof. Do you think that's a good idea?

In general, I don't think that's a good idea because if we, well, we, remember the time when mining happened mainly on GPUs. And one problem that we saw at that time was that malware could take over people's computers. And for example, botnets were controlling a huge amount of mining power, hash power. And that was certainly not good. ASIC has helped that. And another reason, of course, is to have a secure network, you're afraid that there's some entity that could have a technology advantage.

And once we have ASICs on the smallest process node available, we're pretty sure that no entity can gain such an advantage. I think if you do, sorry, if you do ASIC resistant, as they call it, literally you're moving it off of these specialized ASIC processors onto NVDA or AMD or Intel processor. So essentially, you are still hostage to something, except that now the revenue has moved away from the Bitcoin ecosystem into the semiconductor ecosystem. And that is essentially that has happened. And secondly, I think the network will be less secure because you have to use a lot more of those processors,

which are more cap expensive in order to do what we are doing. So, I mean, we can do with one ASIC, which I think would take something like 800 CPUs to do. So I mean, so I think the way we are doing it is we are making the network secure at a much lower cost. tremendously less wasteful power. And in Kepax. There's a focus on that. I mean, we want to be more and more power efficient. So that's a good segue.

So with Bitcoin mining being dominated by the cost of power and the price of Bitcoin, what is the trend? What is happening to improve that situation? Yeah, to add to that, I think this is the interesting thing. We're seeing this build up, as Dave was saying as well. the network is growing over and overall. And as the capacity grows, so does the cost of mining. It starts to shift, not just from investing in the servers and the chips,

but actually running these large, kind of this large infrastructure and the cost that comes with it. I think, definitely a bit fewer. One of the things that we're really kind of focused on is two things. How can we actually contribute to generating our own kind of electricity over time, continuing to drive the cost down of that? And secondly, how do we make sure that we actually do something you do something of use or recycle the energy we're using.

So kind of as we start looking at the immersion cooling, how can we recycle some of the energy? And even thinking more creatively, like if are there any uses in people's home where they might use heating anyway, can we come with creative products there that actually helps both the distributed nature of the Bitcoin network, as well as kind of make something useful as from the side product that we're creating with the kind of processing, not mining. I agree with you.

So, actually, I would say that finally, Bitcoin, Essex are attracting the right kind of people into the industry, which is physicists and semiconductor specialists who are saying, wow, we have all these ideas that we wanted to try in the microprocessors, but they never worked because the microprocessors are too complicated. All those wave pipes. See, there are many, many techniques in which you can cut down power by an order of 10x. but those ideas never worked in microprocessor and other complex chips but they are perfect for A6 that's what we are doing today. Why? Because these A6 are very simple pipeline structures

that we can apply a lot of these physics-oriented ideas. I think in the next two years we are going to see a dramatic improvement in these A6, not driven by the folks that we have building A6 today, but these are the physicists and semiconductor professionals who so far we were not able to attract. But now we are attracting. And I think in the next year or so, we will see dramatic improvement in power and CAPEX. I actually think that by the end of next year, we should see 0.1 watt per gigahash kind of machines out there.

By the end of next year. Yes. I agree with that. But I think that that's where the leveling starts to be seen in the network growth. I think the network growth can simply not be sustained in this vertical singularity style. chart that we see right now in network growth. But that's okay. I think A6 chips allow us to

scale the network out and secure it effectively and efficiently. And to build on that, like we know we're in this build-ups kind of phase, right? And there's going to be a point of equilibrium where actually this is where the real competition between kind of us developing on the A6 side starts to happen, right? The cost curve

is not going to be the same for everyone. And those that are able to come out with the best chips will be able to basically continue to add capacity to network in a way that's still profitable. While others that are unable to do that, not just on the chip level, but increasingly also the data center operations started to become very important. It's going to make a big difference. And there's going to be a big shift, I think, between companies that are able to keep at the leading edge and those that are kind of falling behind. And the good thing about Bitcoin Asix is that they allow you to use your creativity. to innovate and to differentiate as well as the process node. So what kind of things are happening in that area?

Anyone want to? So I think although we use the term ASIC, we use that to mean that the innovation is happening at the ASIC level, at the server level and at the data center level, right? So when we are building our machines, we build at the data center level. When you are building at the data center level, you are looking at what happens in an aisle. For example, when the aisle is cooled down,

there is a thermal gradient from the start of the aisle to the end of the aisle. There's a thermal gradient from the bottom of the aisle to the top of the aisle. So if you apply uniformly the same machine across the aisle, you have a different thermal profile as opposed to different machines which can be tuned to their location

where are there in the aisle. So this is, I'm just explaining about several things we are looking at, is that we can tune the machine to its location in the aisle and then reduce the power and optimize the performance. So what I want to say is that as we are going forward, we have now operated 10,000 machines, 50,000 machines. Now we have to figure out how to run 100,000, 300,000 machines.

We have to go to virtualize machines. So I think virtualization software will take very important. We should be able to run entire mine or transaction processing center from a laptop. Be able to change the temperature performance. of a single ASIC or a core in a single ASIC, of an aisle to any particular server in the aisle. So I think software will also become a very important aspect. So hardware will start evolving in the data center style

and software will start in the virtualized direction. We have that now, the heat optimization. We use Post-it notes on the racks to keep track of the temperature. Always the low cost. That's the equivalent of what we decided. So I spent a lot of my career in the telecoms industry, right? And you have these big network operating centers where you can monitor all things in the infrastructure. I think this is the phase that we're now entering with mining as well,

where this becomes a very kind of centralized thing with a lot of flexibility, the ability to switch things on and off to move from one pull to the other and kind of create that kind of flexibility and kind of insurance of operational efficiency. So in the past, Bitcoin mining hardware has been available for sale to home miners because it's been packaged in, you know, in computer systems. sized boxes, but there seems to be a trend towards building for the data center. What happens to the home miner in that situation? I think it's a win for the home miner because previously home miner had to buy a machine

and then, which only comes in certain sizes, then they have to operate it, and it depends on the electricity charge that in this area, right? Now, if you go to a data center concept and to cloud mining, cloud hashing concept, they have now ability to operate their piece of the hardware because of fractional sale at a very low cost. And so I don't think it discourages home miner. I think it fractionalization of these data centers will allow just like how Amazon Web Services has not discouraged people from working on data centers. It has encouraged hundreds of thousands of businesses. I think this cloud mining, cloud hashing kind of concept, which will allow people to have cloud hashing at a very low cost,

should encourage people to invest whatever amount of money they have into the transition processing. At the moment, I think the economy of scale just tends to favor data center installations at the moment because they are more cost effective. So there's certainly ways despite this fact to keep the decentralization up, because it's, for example, possible to keep the ownership of the chip by cryptographic means in the hands of a home miner and have the chip only accept to work on problems that are actually signed by the home miner, for example. In this way, you can have your chip hosted by somebody else and still have full control over what it's hashing on. That is certainly a way to combine the concept of centralized data center and decentralized ownership.

So, to add to that, I think there's been a move from great, getting the best hardware to officially kind of basically getting not only the best hardware, by getting the best data center location, getting the best electricity contract. And I think that's a progression that's increasingly hard for kind of a home miner to do, right, if you're a rational home miner. So there are two things like it is indeed kind of a progression towards managed services contracts for large miners, cloud hatching. Unfortunately, a lot of the cloud hatching contracts nowadays aren't really.

economically feasible if you kind of to take the calculations on it. The other thing is like, and Dave and I were talking about, it's like there's an element of perhaps non-rational miners. A lot of people would love to be involved and add to the network, right? So I expect, and something that we're thinking about, can we come up with consumer-oriented products in the future that may not necessarily have a great return on it, but be great for people to feel that are contributing to the overall network.

Yeah, there's definitely, that home miners are definitely interested in being part of the network. more than getting a return on their investment. We're finding that a step up from the home miner is the small business entrepreneur who wants to get into mining. So part of our focus on empowering other miners to start up out there is to identify these small startups that actually have favorable economics. They may not have the power prices that I pay, but they have maybe local subsidy from

their local municipality. No one has the power price that you pay. But there are economics out there that actually justify mining where I didn't expect to find them. So we're seeing that there is an opportunity for, granted, it's not home mining, but it's sort of the next level above. It's the small business startup mine that has the ability to make the numbers work. So this year there's been a lot of innovation and creativity and variation among the different A6 from different technology companies, but there's an inevitable harmonization as the

innovation and the tricks that can be done to improve performance become the same. And then the process knows is a limit to where we can go on that. So what's going to happen in the future? I think in two generations, nine nanometer beyond, I think we should be on Mood's Law. So I think till that time, because of the very nature of the SHA-2-56, we should be able to apply other techniques that were not found useful in the normal microprocessor and the chip design. So we have maybe two generations until we hit the Mood's law, but I think those specialized design

tricks will be useful in the next two generations. I think the interesting thing as well, if you would actually do a comparison of ASICs today, There's actually a wide variety of performance matters. So, Bitfure, I think we're very proud that kind of R-55 nanometers is actually at performance doing a lot better than some of the competitors at 28, right? And rather of that has to do with kind of having a bespoke approach to some of these things, and while a lot of others are using kind of pre-configured components. And I think there's a lot, a lot still to do as a way of kind of driving over performance. So I think we'll definitely have a long, long range still where we'll see.

performance improvement that go beyond kind of Morse law that people know from kind of general GPUs and there'll be a little bit... You see that happening over next three generations or you see that over... So what I was saying is that I see that happening, but I think inevitably in two generations you will get into the same situation that microprocessor business and others eventually all the tricks play out and then everybody learns all the tricks. And then basically you come onto the Moors Law and then you go process notes to process Right now, there's been a very low take-up in the smaller geometry processes,

and K&C minor is the first one to use 20 nanometer, and no one else has announced that yet and beyond. The smaller geometries are more expensive, and so there are different trade-offs. It comes down to the financial incentive. So Bitcoin prices shoot up to 2000. You're going to see people going after smaller nodes. It's going to accelerate that going after those tricks that Nandi is talking about.

So actually, from our point of view, we are already working on a smaller node because we believe that there has to be, our customers are buying equipment from us and investing in us, not because we have a machine today, because we have a roadmap. So we have shared a roadmap with them on which there is a 16 nanometer machine and there is a 9 nanometer machine. So we have to constantly have funds available and on our roadmap to invest in that. So it's like any other semiconductor company that has to have that kind of a CAPEX. There are two ways to achieve low power. I mean, there's the go jump to the next node, which is expensive but quick, or there's the spend more time optimizing the full custom. And in the end, like, anyone who's going to make an investment decision to invest in mining, right?

They're going to look at the total cost of ownership, right? They're not just going to look at what's the price X or Y. So they're going to look at what is actually my price per gig a hash, dollar per gigash and whatever. and what I'm going to pay in electricity over a certain period of time, right? So it's trying to optimize across those. That's a great prompt. So what is the relative capital cost of buying the hardware versus the operating costs of running it for, say, a year?

What's the balance? Let me give you examples. Today, I think it will cost you around $2 million to buy one petahash. And it will cost you around one point. I promise you we can buy it cheaper from this guy over here today. Sure, sure, sure. You manufacture yourself. Yeah, no, no, I'm just buying.

I'll talk about manufacturing in a moment. Let's say it's $2 million. Let's call a number between $1.5 to $2 million. I think it covers the range, right? I think I'm covering his range now. So $1.5 to $2 million to buy something, and I think it will cost you approximately $0.8 to $1.2 million to run it for a year. These are the two numbers.

So you're talking approximately $50 or thereabouts for a little bit more to buy than to run. Sorry. Dave, are you seeing the same thing? I think the way I've been looking at it recently, I guess it's easier to talk about right now because it's in the forefront of my mind is, over the next two years, there's going to be 2.6 million Bitcoin produced. At current prices, that's almost $2 billion market opportunity to go after. So the question is, is how much is it cost to invest to capture a large market share of that $2 billion opportunity? It's far less than $2 billion. you could invest

$100 million and capture a significant share of that $2 billion. That's kind of how I've been looking at it. So the business model of the mining industry has been undergoing massive change lately. And customers are no longer happy doing pre-orders. AIC companies are no longer happy doing pre-orders because customers arbitraise the price

of Bitcoin versus their pre-order and cancel and so on. So what's going to happen to the business model of mining? Yeah, I think gone of those days where you could use the pre-order to fund your AC development. So I think which means that the people who will be left standing are the ones who have either already created those funds from their previous mining operations to invest or they have ability to raise funds to invest. So because if you cannot raise the money through the customers, the only way you have is these two teams I've managed. Yeah, the irony with that is that the home miner that, the employer that complains, about decentralization is actually contributing to the centralization by not ordering hardware. That's right.

So in general, I think the pre-order days are gone, right? But I also think that a lot depends on the Bitcoin economy and especially on the Bitcoin price in general. I think the whole market can turn quite quickly if we see another price rise. Because that will also make home miners more competitive again. And home miners can react more quickly to this. In data center deployment, there's a lot of inertia. If you want to react to an increase in Bitcoin price by deploying more hardware, you can not do it instantly because we're talking about multi-megawatt installations.

And they used to take several months at least. Well, yeah. So I think this is one of the competing things going in the future as well. I think you'll see some announcement of us shortly of new daddy centers that were opening. And I think increasingly it's not just competing on designing the ASICs, but also how quickly can you add capacity. And I think that's going to be an increasing differentiator, again, between miners and operators helping them.

And the biggest deal, I mean, the last year, mining ASIC companies effectively crowdfunded their costs, which are in the millions of NREs by pre-orders or large customers, and that's gone. So now... We organically grew. We did some initial retail sales, but then we organically grew everything else to this point. But now we're reaching this point where we're recognizing that if we're going to scale along with the network and try to hold on to our market share, we have to find a way to get out to market quicker. And our way of doing that is to decentralize our industrial mining by teaching others how to do it and then remove the uncertainty of the pre-order problem by sending them the hardware with no up front costs.

Yeah, very novel. That franchise approach is very novel. Yeah, and I think like Bitfure is always taking the choice to not do that kind of pre-funding approach. And I think it's deliberate. I think a lot of companies that have gone that route have actually ended up with significant problems. It's very dangerous.

It is very dangerous, and it's actually bad for the reputation of Bitcoin overall. The interesting thing is it does indeed kind of lead to, again, kind of an opportunity for separation in the sense that the amounts that need to be pre-financed to kind of roll out ever bigger volumes are going to grow, right? So I think it's incredibly important

for specifically those in the ASIC production to be very well funded and it's not always easy for everyone to get that. Now the stakes are so high, does that mean there are no new ASIC entrance? Because the only ones that are around now are the ones that either had a previous generation and raised the money or mined to Bitcoin's

or that are VC funded, of which there's very few of those. Yeah, I think there is a fairly big barrier to entry now. Yes. Yeah. There's a number of barriers to entry. There's certainly a lot of capital required and simply creating the IP also takes a while and the IP is actually the least expensive part of creating an ASIC, isn't it? Right. But it takes time to develop it and to catch up with the existing manufacturers is a barrier to entry. I think the other thing is a customer acquisition. I think previously with a crowdfunding model, you can open a website and

sell to thousands of customers. That's one thing. Now the customers are more larger customers. I think one of the deals we announced recently at three, four days ago, that somebody, one of our customers has bought a nine petahash system from us. Now, developing these relationships is a very expensive venture, right? So for somebody to have just capital to ASIC is not good enough anymore, is to have these relationship with different customers. It takes a long time, which also very expensive, who will buy these multiple PataHash systems from you. And you have to have these customers.

Why? Because if you're spending $5, 7, $8 million in ASIC, a whole server development, you cannot have people buying you half a Pata Hash, two Pata Hage, it won't amortize your cost. So I think customer acquisition cost is also very expensive. So let's do one more question before we open it up to the floor. I want to hear a number from each of you. Today, we're on about 140 petahash on the Bitcoin mining network. What's your prediction for the last day of this year?

No, no, 140 was the start at the panel. Let's ask them how much more it has been since started the panel. Since we started this conversation. Okay, so let's start. Okay, Timo, what's your prediction for end of the year, last day of the year? So I have no better guess than to extrapolate the current growth of network, which is 1% daily, and that will put us at 600. Navajo, what do you think?

I think it's between 350 and 400. Dave? I think it slows down and we make more than 300. We make maybe 350, 400. Yeah, I think it's just fascinating. If you look at the end of last year, if you look at how much the network kind of multiplied over like a 90-day period. At the end of last year, it was around but 11 times over 90-day period, right?

That's come down this year gradually to, every 90 days or so, the kind of the network capacity triples. But it's gradually coming down. So some of the scenarios we've done kind of brings us between, kind of 500 to 700. I'm in the 700 camp myself. Yeah, I'm probably kind of, we took a bet. It's always tricky to the program.

The point is what other constraints are there, right? It isn't just the production, but it's also, yeah, and it's also like the data center. The investment in power infrastructure, and is that incentivized by the current price? Granted, if price goes up next week considerably, there will be investment funds available for the capital. If we were to believe the 700, which means, let's say today is 150, 550 at 1.5 million per petahash, I think which basically will have to assume that somewhere in the world, there is 780 to 800 billion, billion in CAPEX going into... Going in plus that much into the power.

So we are talking about... Dave alone is going to do that. So I think it's $1.2 billion. I think there is $1.2 billion. I think that's possible. I don't necessarily think there's that much of that regard. But I think just give you a number, I think it's going to be over 600. But the interesting thing is what are the constraints, right?

I think it's increasingly on the data center side rather than the production. Great. Thank you, everyone. We can open it up to questions from the floor. Anyone have any questions for our, oh, there we go. Microphone coming your way. This is a little bit more with regards to pool mining. But considering this sort of a strange situation where the closer your approach,

51%, the more you are potentially damaging Bitcoin, yet you have a profit incentive to try and capture as much of the mine Bitcoin, do you potentially see in the future, perhaps some sort of cartel where large mining pools or groups come to an agreement to have a steady percent growth but sort of stay where they are in terms of percentage or someone's sort of other agreement to sort of avoid this strange sort of economic situation where there's more that you can make. I think that we'll always be in competition. So there'll be sort of this cooperatition where.

We'll agree to be responsible in the way that we grow. For example, we can decentralize pools out in a way that doesn't give one pool operator or one pool entity too much power. But at the end of the day, we're going to battle. I mean, to add to that, I think, so sometimes I feel that people confusing this issue a little bit. Because it's in nobody's interest to get it to 51% as such, right? Everybody knows that. So everybody here has kind of money at stake to make sure the credibility of the network stays intact. So in Bitfury's case, like when we saw one of the pools that we kind of participated in, kind of reaching that level, we made a decision to switch some of that capacity elsewhere.

And we did that deliberately. And as I know that others have done as well with the aim of kind of protecting the integrity. And that is just selfish mining. It's kind of optimizing people's interest and it works. So I think that will continue to happen. Great. Is that it or do we do one more question? Well, it's up to you really.

Is there one more urgent question, burning question? Yeah. There is one here, right here. Yes, I think in general people are brainwashed about 51% attacks because, you know, it's not about owning capacity or it's not in anybody interest to, like you say, you know. But, you know, it's only about having this hash power for, you know, it's not about having this hash power for. like one hour. And we have heard here a lot about cloud mining, but cloud mining precisely facilitates that process of acquiring hash power from somebody in a very short time frame. So any comments about this?

I don't know how to synthesize this question. Anyone ever? There'll be, I think there'll be kind of there's already at this form, kind of at this panel, right, there's enough competition to kind of keep that going. And I'm also, I'm pretty confident there will be solutions to also kind of make a distinction over time. And I know there are a lot of people thinking about this to not make, kind of the pool being the determinant of whether or not a 51% attack can take place. So you can even have a split at a data center level of who actually controls which service contributing to

which pool, right? Or even use multi-sig approaches for some. And even with cloud hashing, the gigahashers could be pointed at a different pool. even in the same network. So you can certainly avoid the 51% problem. Cloud hatching itself doesn't contribute to that problem. It's more a way of consolidating hardware into one place to efficiently mine rather than where power is cheap

and cooling is efficient and that kind of thing and then selling the contracts to the end users instead of them owning the box themselves. And when they own the box themselves, it's a physical box. has to go somewhere, it's quite big, it's noisy, it pumps out heat. I mean, the modern mining hardware isn't that great in the home and is actually better seated to the data center.

And that being said, I am very hopeful that we have a lot of smart people in Bitcoin world and we will find a technical solution to the 51% and yeah. Pierre Nova has made a public statement that we will fund or partially fund with other people these developers working on this solution. So certainly we will find our technical solution and I think we should all hope and I think a whole bunch of technical people

are working on it. And if you were at the talk earlier today, my previous panel, we said there's a short-term solution which is convincing the guilty parties to behave more responsibly and educating the community to point their mining power

in a more decentralized way, but long-term, we'll solve it with the technical solution. So I think we'll be fine. And I think if regulation comes to mining, especially in the U.S., that's one of the things I think will be one of the center points that will be discussed as either to avoid 51% or if you can demonstrate that you're responsibly putting up more than that, that you can demonstrate that and that's controlled and insured by a regulatory agency. I'd like to thank our mining panel.

Thank you very much, everyone. Thank you. Thanks. Thanks so much for listening to our coverage of Coin Summit. If you enjoyed this episode, please support us with your donation. It really helps us traveling to conferences and produce high quality content for you. You could donate at episode of Bitcoin.com slash tips, where we have our tipping addresses and also an option for donation subscriptions.

Your support is much appreciated and special thanks to those generous souls we have already donated to the podcast. Thank you.