The inspiration of this article comes from spending a weekend with Nick Foster.
Over the past few years, there has been an incredible amount of interest dedicated toward the bitcoin mining space. It is apparent that many interested parties lack the understanding of how difficult it is to be a participant in this industry, and the amount of work that goes into plugging in a machine so that it begins hashing. This lack of understanding can be attributed to a general disconnect from the complexities of the entire process, because you cannot truly understand it until you try it. What I would like to convey in this article is how much work is necessary in order to build the infrastructure needed to process a Bitcoin transaction.
To be successful in this industry, you need to have some sort of edge over the competition. One such edge is being a maniac. Just passively strolling onto the scene with a “How hard can plugging in computers be?” attitude will not cut it. Having access to capital is not a guarantee that you will make it.
This market is incredibly difficult not only from a Bitcoin-economic standpoint, but also a manpower, hardware, regulatory and logistical standpoint. Difficulties have been massively exacerbated by current supply-chain and manufacturing complexities. Mining is not for the faint of heart, and to be successful in the business you have to be relentless.
Proof Of Work
Bitcoin wants the highest possible amount of effort expended to create the most security for the network. The network needs this to counter bad actors. That being said, increasing network hash rate is a tremendous amount of work just in and of itself.
A petahash consists of about 10 S19s and an exahash is about 10,000 S19s. So, looking at a network hashrate of 200 EH, that means — translated in S19s as a measurement — that there is an equivalent of two million S19s currently plugged in. 312 S19s comes out to a megawatt, so that means that the entire network consumes roughly 6.4 terawatts of power. In reality, a large portion of ASICs currently in circulation are not nearly as efficient as the S19. That means my estimates on power consumption are extremely low.
The math done here is a massive oversimplification. An S19 has a power consumption of around 3.2 kW and has an efficiency of around 29.5 watts per terahash (W/Th). S9s in comparison are around 85 W/Th. The University of Cambridge’s Centre for Alternative Finance did a much more in-depth attempt to measure Bitcoin’s energy usage which you can check out here: Cambridge Bitcoin Electricity Consumption Index (CBECI). In comparison to global energy usage, Bitcoin mining is a rounding error, but even still it is quite impressive.
The top-10 mining pools have roughly 191 EH directed between them. They mine the majority of blocks. That means for Slushpool (9 EH), it takes the work of almost 100,000 S19s. The amount of work that goes behind a user being able to send a Bitcoin transaction is absolutely astronomical. In the rest of this article, I want to talk about the proof of work a miner has to show in setting up their machines in order to make this all possible.
There Are No Experts
The mining industry is constantly changing. When an ASIC manufacturer releases a new machine, everything changes. Even as they continue to produce the machine, things are constantly changing as they tweak it. Manufacturers are often not forthcoming with changes they make, so users have to work with trying to piece together information from equipment they have.
By the time you get a grasp on a certain topic, everything will have changed. To be successful, individuals need to constantly be on their toes, willing to experiment through trial and error. Individuals also need to have connections throughout the industry to source accurate information in order to be successful. Bitcoin mining may be the most competitive industry on Earth right now, so individuals are often not too transparent with the data they are sharing. This creates an incredibly confusing landscape to wade through.
ASIC trends seem to be moving towards increasing density of hash rate and power consumption. For example, the Antminer S9 has a power consumption of 1,400W; The Antminer S19 has a power consumption of 3,250W; The Whatsminer M53 has a power consumption of 6,554W. The massive increase in power consumption means that typical electrical infrastructure changes from generation to generation of ASIC. Infrastructure and methods that worked in the past will most likely not work in the future. It takes constant diligence and work to stay updated on trends.
There is a big reason why Kaboomracks is not only in business but is thriving. Logistics are incredibly difficult and are only getting more difficult. Our employees for some reason really enjoy pain and are willing to run headlong into the mess which is international logistics. Even if you have been in the space for a while, the amount of effort needed to get things from point A to point B is astronomical. First, you have to find what you need. Then, you have to verify that the vendor you are working with is legitimate and is not going to scam you. Next, you have to determine how you are going to get it where you need to. All of that is a tremendous amount of work. Knowing when and why you may want to send something by air versus sea is a big decision that needs to be calculated and takes time to determine. What do you do if you do everything right but the guy with the shipping company is having a bad day and decides to run a forklift through your pallet? These are things that can only be solved through experience and an immense amount of time, research, yelling and pulling your hair out.
The hardware market is incredibly difficult for a variety of reasons. There are a ton of calculations you have to make in the process to ensure you are successful. There are a variety of ASIC suppliers, but which ones will actually deliver you a unit that will get you where you need to go? How many power supplies and control boards should you have on hand for inevitable failures? What amount of downtime is tolerable for you to be willing to send your machine for warranty repair? These are not simple questions to answer and vary depending on the machine and which production run they were manufactured in. These decisions also depend on where you live and what the manufacturer warranty process looks like.
For example, many people got completely turned upside down by Bitmain’s S17 and its incredibly high failure rate as a result of heat-sink issues. That was something that was impossible to know for individuals who preordered it. Knowing when and how to enter into buying ASICs is also incredibly difficult because you are having to time the bitcoin price. ASIC prices act like a lever and the bitcoin price is a fulcrum. If the bitcoin price shoots up, ASIC prices will shoot up even more. Knowing the warranty game is a challenge because things are constantly changing.
If you live in Canada, you cannot send units for warranty repair in the U.S. This is because there are tariffs preventing it from being cost-effective and there are no certified Bitmain warranty centers in Canada. They expect you to send your unit all the way to Hong Kong, which is absurdly expensive from a shipping standpoint, but also a time standpoint. In the same category, if you have a PSU go bad and you live anywhere in the world, the only certified Bitmain warranty center that will accept it is located in Hong Kong as well.
This means that you as an operator have to make a decision on the front-end on what equipment you will buy depending on expected failure rate, with almost no information. You have to have a plan on what you will do when the units fail. This being said, this is hard work and will not be solved in a day. It will be informed by data that you collect from your own experience, but also anecdotes from other individuals in the industry.
McDonalds, Target and other major businesses are having an incredibly difficult time finding people to work for them. The mining industry is not an industry you can just throw bodies at. You need to find the right people, which is practically impossible. You may find some solid individuals with almost no experience in the field, but then you have to train them which takes time and energy that you may not have available. It’s difficult to know if in three months, an individual will be able to effectively troubleshoot a machine, deploy miner management software, tune aftermarket firmware or understand the intricacies of power consumption.
There are many disciplines an individual has to understand, which is made incredibly difficult by the fact that there is almost no documentation on these topics for you to pull from. As an operator, you have to be an expert by having gotten experience to be able to answer this. It is possible to manage with little experience if you have connections with individuals in the field you have to pull from. The moral of the story here is that this is work and you cannot just throw a random person into the fray and expect it to work out immediately.
Finding the right individual is made more difficult by the fact that mining is highly dependent on location. You cannot find cheap power everywhere. Oftentimes you have to pull from individuals in your general area which really limits those available with the skill set to accomplish what you need them to.
Having a background in Bitcoin does not guarantee that the individual is a good fit for mining. Likewise, having a background in electrical engineering or hardware is also not a guarantee that an individual will be a good fit. Individuals coming into the space must have the capacity to be constantly learning and dealing with painful headaches. That being said, to find an individual like this is a tremendous amount of work. It is also a tremendous amount of time and work to develop them right, in order for them to be successful.
Retail hosting and large-scale hosting are two entirely different things. Most large-scale hosts deal in megawatts and a larger number of units. That means that even to be considered as a potential client, you need a minimum of 300 new generation miners. Finding the right host, with the right power rates that will be responsive to your needs is difficult and takes knowledge of the space. Energy markets are in turmoil, which can make working with a host incredibly challenging right now. Hosts face the same risks you do in that you may build up a ton of infrastructure only to get rug pulled by regulations or your electric companies.
Hosting is definitely advantageous for the new operator, as they do not have to wade through as much unknown, as described above. Hosting gives you the ability to outsource a lot of the headaches of actually deploying and maintaining equipment, but it also gives you less control over your miners. Typically, the host will include a hosting fee in the electric rate, and you will have to determine if this makes sense for you as an operator, versus deploying your own infrastructure and hiring individuals to maintain it.
Doing the correct due diligence on the host is a tremendous amount of work. You cannot be too thorough and should not trust them just because someone said good things about them or they had good marketing. You need to be confident that your host will not get rug pulled by their power company or will be unable to service your machines and keep them online.
If you go and plug your miner information into a mining calculator, odds are that profitability will look incredibly different a month later. This makes it difficult to determine the future profitability of your machines. When you are pricing things out, there are two factors which can turn your calculation upside down really fast.
The first one is the bitcoin price. If it plummets, you will still be mining the same amount of bitcoin, but essentially your electric rates, compared to your income, increased dramatically. You have to be prepared to watch your profits shrink incredibly quickly. If it becomes unprofitable for a lot of miners, they will have to unplug and there will be a difficulty adjustment. As difficulty drops, as a result of fewer participants, your mining rewards will increase in bitcoin terms.
The second factor that can impact your profitability is difficulty increasing. On average, hash rate has doubled every year. As hash rate increases, the amount of bitcoin you mine will go down. This typically is okay because the dollar price goes up over time, but in times where price is down and hash rate is climbing, it can become incredibly painful for you. One circumstance that leads to this happening is the release of a new generation of hardware.
Keeping these things in mind, it is important to really weigh the up-front costs you are willing to pay, as it will add to your return-on-investment (ROI) time. Most people get into mining when times are good and are unaware of how ugly things can get when times are rough. Doing research and talking with people who have been in the space can definitely save you from heartache.
Power Generation And Procurement
There are multiple ways for an operator to acquire their power. By far, most individuals acquire it by buying it from the grid. A smaller, hard-core group of masochists choose to generate their own power by setting up generators on oil and natural gas wells. Either way, there is a tremendous amount of energy that needs to be expended to be able to find inexpensive energy.
A lot of people are concerned with renewables. As a miner, you have to decide what energy sources are reliable and consistent, rather than what is the fad. This takes real, hard math to determine. There are definitely different models in regard to types of power. Some people find it attractive to control their power production, and others prefer to just buy it from a producer.
Energy markets are tough, and it is not easy to just go to your local power company and negotiate favorable prices. Oftentimes, you have to do some digging to find a location next to a substation and try to determine how much excess power there is. There are firms that you can hire to find power for you. Another option is to hang around substations and pass out $100 bills and beer to the workers servicing it to try and get some inside information.
One of the best ways to find excess electrical capacity is to look for industrial places where a lot of large power consumption businesses have moved from or shut down. Whichever direction you go, it will take time and effort to figure this out before you break ground on your site.
Lead times on transformers are absurd right now; The lead times on just about anything is horrible right now. First you have to get the equipment you need, then you need to find an electrician to install it. Finding an electrician is not just as simple as calling one up. You need to find someone who is competent enough to do the job correctly, but also is available to do it in a timely manner. In a time when everyone is understaffed — and booked up — this can be quite the task.
When you take shortcuts, you will most likely run into issues. Timelines are incredibly difficult to stick to or formulate because there are so many moving pieces that can disrupt your goals. All this being said, this is more work.
Plugging in ASICs and installing infrastructure is dangerous. It is a completely different business than traditional server hosting or what you would do in a data center. Oftentimes, operators have to improvise and develop the products they need for themselves with the manufacturer. Having a smart power distribution unit (PDU) that collects power consumption data is incredibly useful, but having it made for the type of power consumption that ASICs need is a whole different story. Learning that this process is constantly evolving is a lot of work.
Aftermarket firmware is another discussion to have as it can increase your hash rate/power consumption, without up-front equipment costs. If you are doing immersion, it is almost a must to run firmware for overclocking. Firmware adds a whole new world of possibilities but also headaches.
There are some challenges that aftermarket firmware creates. It adds complexity to your operation and can increase machine downtime as a result. Also, if you are not aware of the specs on your equipment, you may damage your electrical infrastructure from increasing the power consumption, or even damage the ASICs themselves. Some firmware is not compatible with all control boards. If you do decide to implement it, it may mean swapping a bunch of control boards. To get the most out of overclocking, it will mean potentially swapping all the PSUs on the machines.
Another challenge is learning how to properly use the firmware. Tuning specifications will be different between machine models: Some models are better for overclocking than others. You can outsource this or take the time and energy to have someone trained in-house.
All in all, some find pursuing aftermarket firmware incredibly fruitful, but there are still a lot of scenarios you have to consider before using it. Again, something as simple as installing a program to run on your machines is a ton of work to implement and can have a bunch of consequences.
Immersion opens up the door to a lot of possibilities. It is advantageous, especially in hotter climates, and will extend the life cycle of your machines. You can dramatically increase the overclocking using firmware. Immersion adds a massive increase to your up-front cost, but you may find it worth pursuing. There are many advantages to using it — but once again, it adds a lot of complexity to your operation.
Machines will run substantially more efficiently in immersion, which makes a massive difference at scale. Sourcing the proper equipment for immersion adds to the cluster, which is dealing with long lead times and logistics. There are a lot of people making tanks out there, which makes it difficult to determine what is good quality.
Another difficulty that immersion creates is making servicing machines more challenging. If a power supply unit (PSU) or a control board goes bad, there is a lot more work needed to pull the machine out of the oil and replace the part. Even prepping machines for immersion in the first place can be a lot of work.
Payout structures are different from pool to pool. Also, your potential income will be different from pool to pool depending on how many blocks that pool mines. Sometimes, mining pools have bugs and make mistakes, so it is good practice to closely monitor your payouts. Monitoring software like Foreman offers the ability to do this, but costs money.
It is difficult to gather data on this topic because things are constantly changing in regard to pool hash rate and pool infrastructure. What many miners do is set a portion of their hash rate to different pools and gather data. It is good practice to be prepared to switch pools if they have any downtime.
Miner costs are consistent, so miners benefit from having their rewards be consistent. Luck-based rewards systems make much more sense for small miners that want to roll the dice than they do for operations with six-figure electric bills. Understanding payout structures, pool performance and monitoring if your payouts are correct is hard work.
Networking in and of itself is a tremendous task to take on. Having multiple options of ISPs (internet service providers) is important to minimize downtime. It is not usually as simple as setting up with multiple internet providers, as oftentimes they will piggyback off each other and will not be transparent with you about this. This means that if one goes down, several might at the same time.
The physical topology of networking is an in-depth conversation in and of itself. If you yourself are not knowledgeable, it is another thing that can be contracted out. Having control over and being knowledgeable over your network, versus being reliant on outside contractors, is definitely a major advantage. Networking has a tremendously steep learning curve and either way is a tremendous amount of work. Networking for a Bitcoin mine is a lot different than setting up a home router.
Though ASICs do not take up a tremendous amount of bandwidth, they need a good connection with the pool 24/7. Proper networking can help an operator minimize their downtime and improper networking will create a mess. Nothing is simple about this process.
Thanks to Matt Odell’s podcast, Citadel Dispatch, there has been a tremendous amount of interest in mining at home. I shouldn’t give Odell all the credit, as other Twitter personalities such as Diverter_NoKYC, Econoalchemist and Roninminer have been influential as well. That being said, people have proven that you can mine at home and it can make sense.
Once again, the process of setting up ASICs in your home is quite the process. These machines are industrial grade and not made for your home. This is not a flaw in their design, but just how it is. The ASIC market compared to general electronics is tiny, and the home mining market is even smaller. The infrastructure to run multiple machines generally is not there before modifications. For most people, there are only two outlets that can deliver enough power for a S19 or M30s to run: the dryer and oven outlets.
An individual will have to figure out how much power is expendable to them, the rates at which they are charged and how to mitigate the temperature and noise. Doing this properly is important because most individuals do not want to burn down their houses. The biggest challenge for an individual is figuring out how to not burn their house down with limited documentation. All in all, the process of mining at home takes work.
As seen by the most recent China ban, regulatory risk is real. A local or national government oftentimes has no issue with rug pulling an operation. There have been many examples of this and there will be many more going forward. This is a really difficult thing to build a threat model for in our politically volatile environment.
Something to consider on this topic is building out in multiple jurisdictions, versus just one. Having multiple locations allows you to potentially move and build out more if needed, but also means that you need more manpower to manage your sites. It is not easy to just pick up and move, as it is almost starting over from scratch.
Another approach is lobbying local politicians and working really hard to get the buy-in from the local community. Riot has been very effective at this with their Whinstone site in Rockdale, Texas. They are consistently engaging with Bitcoin conferences, media and their local community to get their buy-in. You can often get away with flying under the radar if you are small. When you get bigger, there will be a lot more scrutiny on you. One approach that has been taken is actually setting up local governments with the ability to mine themselves.
Bitcoin mining is ultimately a net benefit to any community, and it is advantageous to teach people about this. Staying in the good graces of the community is probably worth pursuing because it is a shield to protect you but is also worthwhile because it aligns with Bitcoin’s mission of empowering the individual. It is a major opportunity to help strengthen and give back to the community you are operating in. Giving back to the local community is always a good business practice. That being said, all of this is more work added to your plate.
Cheap power does not always mean a safe environment to run a stable mining operation, as seen by Ukraine, Russia and Kazakhstan. When you are choosing the location of your operation, potential geopolitical risks have to be weighed seriously. If a government has shown to be hostile towards mining in the past, they should be taken seriously.
Cross-border sanctions can play a major role. If you are setting up an operation outside of the U.S., something to consider is having your business located outside of the U.S. as well to potentially limit the impact of sanctions. There has been a tremendous amount of growth in U.S.-based mining, but it remains to be seen if the U.S. is truly a safe spot to have a long-term operation.
Virtually all ASICs are produced in China, and the companies producing them are based in China. They have been working around sanctions against China by shipping and moving parts of production out of China. Geopolitical tensions between the U.S. and China may make it far more difficult to source equipment in the future despite this, which may or may not make it advantageous to be outside the U.S. That being said, there is a lot of work needed in order to think through and prepare for a lot of potential scenarios that could have major impacts on your operation.
Are You Built For This?
The oil and gas world is a perfect fit for the mining space, because they are already exposed to and fully understand the pain that is involved with a similar industry. They are also predisposed to understanding energy markets like no one else is. If you are an individual who likes to sit in a nice, air-conditioned office, but does not like to get your hands dirty, mining is not for you.
For some individuals, the industry is too difficult, and they will quickly give up when it does not work. For others, mining is an addiction that you can’t let go of. It consumes your life until all your fingers are bleeding from being cut on fans or pulling out ribbon wires. It is an industry like no other, but it sure as hell is a fun industry to be a part of.
Security For Users
All the effort that goes into proof-of-work makes the network extremely secure. Say for instance, a government or large powerful entity wants to take control of the network. They would have to deploy a monumental amount of equipment. The equipment needed to take over the network just isn’t there and neither are the incentives. The Bitcoin network is backed by thousands of insane masochists chasing a profit to the ends of the Earth.
The media, which represents the incumbent system that Bitcoin is displacing, decries Bitcoin’s energy use. If anything, the complexity and industriousness of this industry should be celebrated. We are building the most secure and best monetary network ever created, which is moving humanity forwards in so many ways. Crazy masochists allow Bitcoin users to move large amounts of money, inexpensively, anywhere in the world without reliance on third parties.
The amount of effort that goes into mining Bitcoin takes an astounding amount of emotional energy. This emotional energy is just as powerful, if not more so, as the financial capital put into setting up an operation. Miners literally bleed going through the process of setting up their mines. Miners will continue to mine, innovate and put pressure on forces trying to attack the network.
There is nothing simple about Bitcoin, and there is absolutely nothing simple about mining bitcoin. The complexity of sourcing and setting up infrastructure is a force that inevitably distributes network hash rate. The average Bitcoin user will never understand the heartache that goes into mining Bitcoin but will still experience the benefits from it.
The competitive and constantly changing nature of the market means that there will be a constant flow of participants entering and exiting. The name of the game is survival. Operators have to be prepared for the worst-case scenario and have to be willing to adjust in accordance with it. There is not an easy path to take in this industry.
To survive, you have to work countless hours in order to make sure that things work. You will hit roadblocks, get massive headaches and wake up in the morning feeling hungover without having taken a single drink. For some strange reason, thousands of us find this kind of work incredibly fulfilling. All of this effort is what it takes to mine bitcoin and is the proof of work behind proof-of-work.
This is a guest post by Kaboomracks Alex. Opinions expressed are entirely their own and do not necessarily reflect those of BTC Inc. or Bitcoin Magazine.