Bitcoin mining – the process of adding new blocks of transactions to the Bitcoin blockchain in exchange for coins – has undergone a transformation in recent years. What began as a niche pursuit for individual cryptocurrency enthusiasts has evolved into an enterprise-scale business, dominated by industrial operators running vast operations.
With the global hashrate surging to record highs and scrutiny on its energy consumption intensifying, miners have moved beyond sheer scale to smarter, more efficient operations.
From predictive maintenance that prevents costly downtime, to automated energy balancing that keeps operations running smoothly, there is potential for artificial intelligence (AI) to help transform mining into a high-tech, data-driven industry.
US miners now account for more than 40% of the global hashrate and will play a central role in the future of crypto mining.
Techopedia spoke with Jared Focose, CEO of the US Division Hosting at GoMining, about how AI can assist in maintenance and power consumption management, how Bitcoin mining can help balance electricity grids, and the importance of transparency and non-fungible tokens (NFTs) in building trust.
Key Takeaways
- Bitcoin mining has shifted from retail hobbyists to large-scale enterprises, requiring more sophisticated tools to stay competitive as the hashrate reaches record highs.
- Bitcoin miners are playing a role in stabilizing power grids, using curtailment to scale back energy use during peak demand and feed power back into the system.
- AI-driven predictive maintenance and energy balancing can potentially enable miners to reduce costs and operate more efficiently in an increasingly resource-intensive industry.
- Transparency and trust are central to the industry – GoMining addresses this by livestreaming data center activity to demonstrate accountability.
US Division at GoMining
Jared Focose is the CEO of the US Division at GoMining. He oversees the company’s operations, strategy, and growth initiatives specific to the US market.
Focose is responsible for driving forward the division’s objectives, ensuring alignment with GoMining’s overall vision and business goals.
The Role of Automation in Predicative Maintenance
Q: What are the types of failures and inefficiencies that can occur in Bitcoin mining, and how do you detect them?
A: In terms of failures, hashboards go out, and we have to change chips. Power supplies and fans go out. Power supplies are probably the most annoying; hashboards are the most common. It depends on the manufacturer and the machine that you’re dealing with.
We buy newer, higher-quality equipment and try to minimize failures. We commit to a 99% uptime; most companies are looking at 90-95%, maybe 97% if they are good at it. We are good at predictive maintenance, which comes from being able to look at all the machines through our monitoring system and understand when one is under-hashing or when it switches off.
We get ahead of it – pull it out, replace the power supply, and put it back – once you have hundreds of thousands of machines, you can build an intuition around it over time. You get the feeling when one is going to fail soon. All that intuition pulls together, along with data that is well-linked.
Q: How does AI-powered predictive maintenance work in a Bitcoin mining environment?
A: The technology still has some maturing to do on that front. We’re working with a couple of different partners to develop a system. We have a monitoring system and a team that manages hundreds of thousands of machines globally at different sites. And it’s still done in a human way.
We haven’t been able to fully integrate AI on top of that because of the level of decision-making required on each individual machine. It is like car diagnostics; there is a reason you can’t just drive it into a shop and plug it in to repair it. They can see some things that are wrong with it, but you need a human to look at it, to understand the whole picture of what it’s supposed to do, and have some experience with that machine specifically.
Layering that has been difficult. The main target is to go bottom up instead of top down, using individual tools and using AI circumstantially to build up a baseline with our infrastructure. Our digital infrastructure is sensitive to big changes, and we want to be careful about what kind of data is going in and out, because of the way the business model operates. Nobody trusts an NFT version of a miner already, which is a physical thing that you’re never going to see.
We have videos on our website where you can see the miners operating in the data center. But there also has to be a transparent data path from the Bitcoin you receive all the way to the miner.
We are very careful about what kind of changes and technologies we implement into that stream, because one, if it’s not broken, don’t fix it, and two, we are going to be liable for whatever happens in that data stream.
“It takes time, and Bitcoin mining is a very physical, very human thing. It takes large teams. You can have a multi-billion-dollar company that’s just got a little asset, but once you tie it to something physical, it’s a risk.
“I give it a year. AI is almost doubling in intelligence every 18 months, so it will get there.”
A: We’re starting with the firmware that connects to our server management system in the data center. This first layer collects data, makes decisions, and predicts things like maintenance needs, power use, and hashrate. The goal is to optimize each machine, then scale that optimization to the whole data center.
The other thing we’re looking at is strategic curtailment and the optimization of our data center power consumption. If we can intelligently understand where the power market is going, instead of turning the whole data center off for curtailment and then turning it back on, we can shut a few machines down and smooth it out.
The power market is a slow-motion market, and a lot of it is predictable. If we can get ahead of that, and we can intelligently manage our consumption, that is ideal for AI to step in and do, but there’s a lot of money on the line, so we have to be careful about how we implement a system like that. An error is very expensive.
Location & Energy Strategy in Bitcoin Mining
Q: What are the considerations when deciding where to locate Bitcoin mining facilities?
A: Push-pull factors usually depend on the location and what we’re trying to accomplish – but most often, it comes down to the cost per kilowatt hour for power. Canada, Alberta, and Texas beat everybody to the punch on that. New Jersey used to breathe cheap power, but now you can go to half the states and figure out a way to get cheap power.
Climate is extremely important – ideally, low humidity and low temperature, but not too cold. It gets challenging once you get really cold, as you have to recirculate the air. Humidity is a more important factor than overall outside temperature, because our cooling systems work 400-500% better. We are using evaporative cooling systems, and you can’t do evaporative if the air is already full of water.
Then it becomes a game of what kind of local political involvement do I need to make sure this is going to be successful, whether the market wants us.
We are putting infrastructure, long-term contracts, and a lot of time and capital into something that is based on a highly volatile market. The level of risk tolerance required to do that needs to be appreciated by the grid that we’re coming into.
If they think that we’re some rich tech guys that are going to come in and make an easy buck sitting on top of their grid, and they’re going to patronizingly give us some power, we don’t want that – other opportunities exist.
Q: How do you approach energy balancing and consumption – are you trying to automate it more to optimize consumption?
A: On a miner level, a lot of people use custom firmware. We don’t; we like to link in with stock firmware. We have good relationships with the manufacturers. We tend to try to stabilize power consumption at the cost of a little bit of optimization. But when you have hundreds of thousands of machines running, you want everything to run flat and smooth.
Our consumption is well understood and well managed, but we do more to manage it on a data center level. We try to avoid curtailments. We can lower our cost of power via curtailment, although that has an impact on the maintenance of the machine. If you shut everything off for a few hours, it is going to cool down, and when you turn it back on, it will heat back up. Those kinds of temperature cycles can be difficult for machines to handle.
But we do take advantage of curtailment for some price impact. It depends on where the machines are located, geographically, what the temperatures are like outside, what the humidity is like, and what kind of cooling we’re running on that system.
Texas is good for curtailment; they give a lot of rewards. We’re finding out that there are other markets that are aggressively trying to compete with ERCOT, the grid in Texas, to provide as low-cost power as possible.
Bitcoin mining has taken the place of batteries – you can put a Bitcoin miner on the grid and ask it to shut down; it acts as a battery because you increase the baseline consumption. Instead of having to peak up and take power from either peaking plants or battery storage, you can reduce the miners’ consumption and bring it back.
The real impact of Bitcoin miners in the world right now – and the reason that people are shaking our hands and bringing us into deals – is because we become a small-town-sized battery that can feed into the grid and stabilize.
Bitcoin mining allows us to take advantage of good power pricing. And grids are getting competitive with each other. They are looking to attract more miners to have the benefit of the grid. They build out substations for us, knowing that we’re going to sign a long-term contract with them, as opposed to in 2017, when trying to get enough power to run 60 S9s, they wouldn’t even take it seriously.
Q: How do you respond to critics who argue that Bitcoin mining is environmentally unsustainable, and what are the ways that mining is changing?
A: We use more sustainable power than any other industry, by a huge margin.
Hydropower, for example, tends to be built on long timelines, and it is built politically and situationally. It is a power of convenience. Niagara Falls, for example, has a dam that was not driven by the necessity of having that power locally available, which drives the cost down.
They have more power than they know what to do with, and they can’t peak that power. They need us around to act as a battery on that part of the grid and give the correct variability of availability of power versus just the massive availability.
Where wind power is installed, they end up bringing Bitcoin mining onto those grids, oftentimes to curtail because the winds can’t produce power during certain times, and they need us to run and then shut down. There is some negotiation that goes on there, and we can help solve some of those problems, but we hope that they have solar or something else nearby to try to fill in the gaps.
I would love to say that we target sustainable energy because we’re awesome and we just love the earth, but realistically, growth and building are driven by economics, and the economics work out with sustainable energy.
Building Trust Through Transparency
Q: Given the issues around trust in Bitcoin mining, how important is transparency?
A: It is our top priority. It is up there along with uptime, maintenance, and expansion into new facilities.
We have live feeds into our data centers in the US, but we’re implementing them at all the data centers that we build. We have to make sure they stay online, and then nobody can say this is a fake or looped feed.
The gap between enterprise and retail is getting bigger, but we tunnel through it with NFTs. We mint and deploy NFTs for customers that give them the rights to a certain number of terahash at a certain efficiency, and those are directly tied to miners in our fleet.
It is hard for people to trust NFTs. Wrestling with the perception that Bitcoin mining and NFTs are a scam has been one of our biggest challenges, and a way to do that is to take and transfer the human experience all the way to the data center level. We rely on our eyes – if you buy this NFT you can put your eyes on the machine that it is connected with, and then you can look at your wallet and see the Bitcoin coming from it, making that loop has enhanced the trust from our community.
We can give retail customers the enterprise-level benefits of mining and the economies of scale that come with it, the reliability that comes with having a huge team, and everything managed well.
The original business model was to tokenize the hashrate and have every 1,000 GoMining tokens equal one terahash for processing.
There are a lot of challenges that come along with that, such as regulation, and the incentives become very difficult to align when the hashrate becomes less valuable over time. NFTs let us give people title to the hashrate forever, as long as you keep up with the efficiency.
An NFT has a certain efficiency, like 20 watts per terahash, and in 20 years, miners might be doing one watt per terahash. But over that time, there is an opportunity to upgrade efficiency as we modernize our fleet.
Innovating the Financial Infrastructure of the Future
Q: As now US miners contribute over 40% of global hashrate, what can other countries learn?
A: The US has a level of baseline infrastructure – the interstate highway system that makes it easy to transport equipment across the country, and the power grid. A focus on long-term infrastructure is helpful, but that takes 20–30 years.
Near term, it is to push innovation in business models. Seeing Bitcoin miners as a battery on the grid has been huge, and this is why we are in Texas. The genesis of that goes back to 2021, when some of the power infrastructure failed, and the peaking plants failed too. Bitcoin miners came into the picture, and that created grid stability. The lesson would be to try to find innovations where you would not imagine they exist.
There is an innovative environment in the US, where people are willing to take crazy ideas and run off a cliff with them and hope they fly. And even huge organizations like ERCOT and multi-billion-dollar companies are able to take these kinds of risks.
I would say, back off the politics and look for smart people trying to do crazy, cool things, and if those people are willing to take the risk on it, figure out a way to help them. Don’t subsidize them if they fail, but don’t stop them if they’re trying to go for something.
Q: How do you see Bitcoin mining progressing from here?
A: It gets more enterprise-level, more efficient, and professional. The suppliers are starting to tighten up quality. We had challenges with that for a long time, but we’re seeing more quality come into the space. The technology grows, the chips and the power efficiency get better.
Eventually, the traditional financial system will be phased out by the technological revolution that is happening on the blockchain side. That means we are trying to build the infrastructure of the future financial system.
We have to make sure that we do it right, all the technology and services are high-grade, the contracts are long-term, and the basis is solid, so that what the world wants to build can be built on top of it.
Bitcoin mining has come a long way from its early days of individual enthusiasts running rigs at home. Today, it’s an industrial-scale operation. From predictive maintenance that keeps machines online to curtailment strategies that help balance power grids, miners are making advances to operate more efficiently, sustainably, and transparently.