The United States Army Should be Mining Bitcoin
Inch by inch, the United States Army is moving towards energy independence. In 2014, the Army established the Office of Energy Initiatives, and they argue that “secure and reliable access to energy and water on Army installations is essential to the Total Army and its ability to deploy, fight, and win in a complex world.” In their quest for secure and reliable energy, the Army has increased its capacity of renewable energies throughout it’s bases. For example, in Georgia, the Army built 3 30 MegaWatt solar farms at Ft. Gordon, Ft. Benning and Ft. Stewart. Additionally, solar energy provides 12% of the total energy used by Ft. Detrick in Maryland and Ft. Hood can produce 62 MW of renewable capacity, split between a 50MW wind farm and 12MW of solar power. These are just a few of the projects that the army has started to increase the reliability of the local power grid, but these projects are just the start. In the long term, the Army will move towards full energy independence from the local power grids. Soon, having a small emergency store of energy won’t be enough for the Army. Current local power grids lack the security and resiliency to fully support the Army’s needs in the event of outages or cyber attacks, and the best way to assure energy resiliency will be through self-controlled power grids.
One of the issues that running a power grid inevitably comes against is the daily energy demand cycles. This issue is compounded with the use of renewable energy sources such as wind and solar where supply levels vary throughout the day. Variable energy sources reduce the reliability of the system at large because of the difference between when energy needs to be used and when energy is created. One potential long term solution is batteries. Unfortunately, the Army is already discovering storing solar energy is not feasible at the scale necessary for an Army post. In addition to the difficulties renewable energy face in providing energy throughout the day, older generation systems can undergo severe wear and tear from being cycled on and off to meet constantly changing energy demands.
Independent power generation will come with its own set of problems that must be answered. Below, I lay out the case for mining Bitcoin instead of cycling power generators on and off throughout the day. This will have several benefits. First will be that the Army will be able to offset the costs of Energy production with rewards from mining Bitcoin. Second will be that power plants will not have to cycle as frequently. This means power generators will last longer and produce less pollution. Lastly, the Army’s power grid will be incentivized to become more robust due to Bitcoin mining incentives that reward cheap electricity. This incentive will force the Army to search for more energy that can be produced cheaply or risk losing out on their investment into Bitcoin mining.
Bitcoin Mining
Bitcoin is an emergent technology that enables the transfer between energy and money. The transformation of energy into money occurs in the process of mining Bitcoin. If Alice wants to send money to Bob, she broadcasts a transaction conveying this information to the Bitcoin network. Miners then hash (put blocks of input through calculations that spit out a string of characters of a fixed length) groups of transactions with varying nonces (random numbers to change the hash) trying to find a hash that contains certain characters at the beginning. When the correct hash is found, a block (list of transactions included by the miner) is added to the Bitcoin blockchain and the process starts over looking for the next block. The characters at the beginning translate to the difficulty level of finding a block, which updates approximately every two weeks to ensure that blocks are found approximately every 10 minutes. In the process of finding blocks, miners include a transaction to themselves that includes the block reward and all transaction fees that were paid towards the block. Currently, the Bitcoin reward is 6.25 Bitcoin along with approximately 0.6 Bitcoin in transaction fees.
Although personal computers can compute hashes, Bitcoin mining has evolved past the point of CPUs being profitable for mining. Today, mining Bitcoin is largely done with devices called Application Specific Integrated Circuits (ASICs). Currently, one of the best ASICs on the market is the Antminer S19 Pro, which computes 110 Terahashes per second (TH/s) while consuming 3250 W of energy. For comparison, the top of the line CPUs can compute around 8kh/s, approximately 10¹¹ less hashes per second. Through keeping power generators on during the day, Bitcoin mining profits will offset and exceed energy costs and reduce wear and tear on power generators.
Methodology
In order to narrow the scope, we are going to take a look at four Army posts, chosen for their diversity in location and operational mission to get a general cross section of the feasibility of mining Bitcoin on Army bases. Ft. Gordon (GA), Ft. Hood (TX), Ft. Detrick (MD) and Ft. Drumm (NY) will be the site of our Bitcoin mining experimentation. From each of these locations we used hourly energy data from the surrounding region provided by the U.S. Energy Information Administration API. Combining this regional data with the amount of energy used by each of these posts throughout the year from the Department of Defense Annual Energy Management and Resilience Report we can model the hourly energy usage of each of these posts. After modeling the energy usage of each of the posts, we calculated how much money Bitcoin mining would be likely to earn each hour, using the below equation.
Finally we’ll sum up the returns for each hour throughout the year to find the overall expected revenue from Bitcoin mining and subtracted the cost of the mining rigs purchased at the beginning of the year. At this point, we can assess the profitability of using different amounts of available energy, looking for the percentage of miners that produce the best returns over the first year. The difference in market conditions between 2020 and 2021 are summarized in the below table.
Before going any further, it is important to acknowledge some of the assumptions this model makes and the validity of these assumptions. The first assumption is that there would be no significant change in market dynamics due to the U.S. Army’s involvement in Bitcoin. This one is tough to gauge because any new entrants throw the delicate balance of the mining market out of whack. As far as the price, it is likely that Bitcoin’s price would appreciate significantly. The first reason is that we have no way of knowing the impact the Army would have on both the price and hashrate market and trying to guess would introduce further error to our modeling. Price, hash-rate and mining rigs available are all inextricably linked and the mining sector is diversified through pools of people around the world that have differing costs and reasons for participating in the market. Additionally, any involvement in Bitcoin by the United States would be viewed extremely bullishly. This allows our model to be closer to the pessimistic view.
Our next assumption is that we assume the Army will buy mining equipment on the open market in order to estimate startup costs. This is another assumption that is likely faulty, but provides a starting point to base costs. The main issue with this assumption is that the Bitcoin mining hardware market is currently dominated by three Chinese manufacturers, which would be an unpalatable source for the United States to source its mining hardware. Purchasing large amounts of Bitcoin mining equipment from Chinese companies would expose the United States Military to a supply chain risk that could prove devastating on numerous fronts. Instead, we assume that the Army would have to invest in some form of infrastructure either internally or through a contractor to buy or manufacture mining rigs. Over time, this should prove to be cheaper by reducing middlemen between the construction of the mining rigs and the Army’s ability to use the hardware. Producing your own mining hardware is not without its risks, however, and would likely incur greater startup costs. This is the price of greater supply chain security.
Finally, although this model is aimed at a future where the Army is producing their own electricity, energy costs were assumed to be market rate in each of the states the posts reside. This is another assumption that is likely high. As previously discussed, the Army has already installed a number of renewable energy sources that create energy regardless of whether it is needed or not. In this instance, the marginal energy that renewable energy sources produce, incurs little additional cost. One issue the Army faces already is that battery technology is not able to store energy for long enough at the scale that the Army would require. In instances such as this, using the energy would be virtually free. Either the Army would mine Bitcoin with the energy, or it would go to waste.
2020 Market Dynamics
The Bitcoin mining market would have been extremely unforgiving to the Army in 2020, according to our model. Although each post would have made between 10 and 30 million in Bitcoin mining earnings, the startup costs make the prospect of mining unappetizing. Two factors likely had the biggest impact on this result. The first is that the Antminer S19 Pro became available in May 2020. Due to its lack of availability earlier in the year, we used the older Antminer S17 Pro. However, other mining corporations would have switched over as soon as possible and Bitmain, the company behind the mining rig, likely switched over even earlier. Additionally, May 2020 saw the third Bitcoin halving. This event results in the mining reward to halve, making the last half of the year extremely competitive for the mining reward at the same time as new equipment is coming online. With 100% of miner utilization, each post would have lost more than 15 million. One notable trend is that after about 75% to 80% miner utilization, profit falls even further. At this point, the mining rigs are not operating long enough to have a significant impact on mining revenue.
2021 Market Dynamics
2021 brought along a new mining rig that would not have been available for all of 2020, and thus was not included in the 2020 analysis. The Antminer S19 Pro more than doubles the hashrate of the Antminer S17, with only a 50% increase in power usage. Coupled with a higher base price to start the year, Bitcoin becomes significantly more profitable to mine.
In fact, based on our model, each of these posts could have paid off the entirety of their energy bill and then some.
This analysis once again showed that mining revenue peaked around 75%-80% of miners utilized. To look more into potential profitability, I used a rate of 75% miner utilization to determine potential earnings at price levels between $0 and $80,000 and hash-rates ranging from 50 million to 250 million TH/s. These graphs show a very clear region where profitability including startup costs approach zero with a fairly large opportunity to cover energy expenses.
Concluding Remarks
Much of the calculations behind the model were done during January 2021 and have made the model look fairly out of date now. The projections based on the initial price of $30,000 and a hashrate of 148.6 million TH/s now stand at about $56,000 and 156.2 respectively, indicating that it has only become more profitable to mine Bitcoin this year. The current market dynamics are incredibly conducive to mining. While these dynamics are ever changing and may not be as favorable in the future, the incentive to increase renewable energy sources for cheaper electricity aligns closely with the Army’s long term goals to become more energy efficient while maintaining and improving robustness within the system. Through incentivizing building cheaper energy stores that can operate at higher levels, the Army’s operational goals can synchronize with the incentives laid out in Bitcoin.
You can follow Orion on twitter at @orionboylston
