In a recent video dialogue hosted by Bitcoin Magazine, Troy Cross, an educator in Philosophy and Humanities at Reed College, investigates the subject of his latest piece for Bitcoin Magazine’s “The Mining Issue,” entitled “Why the Future of Bitcoin Mining is Distributed.” View the complete conversation here.
In the conversation, Troy examines the centralization factors in Bitcoin mining and puts forth a persuasive case for the decentralization of hashrate. Despite the cost efficiencies that have led to the emergence of massive mining operations, he underscores a vital—and potentially financial—necessity for distributing mining authority, providing perspectives on the future of Bitcoin’s ecosystem.
The subsequent article is highlighted in Bitcoin Magazine’s “The Mining Issue”. Subscribe to obtain your copy.
Introduction
When Donald Trump proclaimed his desire for all remaining bitcoin to be “MADE IN THE USA!!!” Bitcoin enthusiasts celebrated. Mining is beneficial, right? We want it to occur here! And indeed, the U.S. is on track to dominate the sector. Publicly traded U.S. miners alone are accountable for 29% of Bitcoin’s hashrate — a figure that appears to be increasing. Pierre Rochard, vice president of research at Riot Platforms, forecasts that by 2028, U.S. miners will generate 60% of the hashrate.
However, let’s be candid: Concentrating a majority of Bitcoin mining in the U.S., particularly in large public mining organizations (instead of having a Bitaxe in every home), is a poor notion. If the bulk of miners exist in a singular country, particularly one as affluent and powerful as the U.S., miner actions would be influenced not only by Satoshi’s meticulously crafted incentives but also by the political motivations of the current administration. Should Trump ever attain his vision, the very future of bitcoin as a non-state currency would be jeopardized.
In the following sections, I describe what a nation-state assault on bitcoin via miner regulations might resemble. Following that, I analyze the incentive mechanisms that have propelled Bitcoin mining into substantial U.S. data facilities controlled by a small number of corporations. Lastly, I argue that the future of Bitcoin mining does not mirror its recent history. In my view, Bitcoin mining will return to a distribution reminiscent of its early days, where miners were as numerous and as geographically scattered as the nodes themselves.
I also propose that notwithstanding some Bitcoiners’ excitement for “hash wars,” and despite political bravado, nation-states actually have a stake in a future where no single country dominates Bitcoin mining. This “non-dominance dynamic” differentiates bitcoin from other technologies, including weaponry, where the incentive to dominate propels nations into competition for market supremacy. Yet with Bitcoin mining, dominating equates to losing. When nation-states grasp this unique game theory, they will assist in safeguarding it against miner consolidation.
The Assault
If the U.S. possessed the majority of hashrate, how could bitcoin be undermined?
With a simple order from the Treasury Department, the U.S. government could instruct miners to blackball certain addresses, for instance, from North Korea or Iran. The government could additionally prohibit miners from constructing on top of chains featuring restricted blocks; in essence, all miners would be barred from appending a block to a chain containing an earlier block with a censored transaction. Major U.S. miners — public entities — would then be obligated to comply with the legislation; executives are averse to incarceration.
Furthermore, even miners located outside the U.S., or private miners within the U.S. opting to disregard the law, would be required to censor. Why? If a renegade miner inserted a blocked transaction into a block, compliant miners would be compelled to orphan that block, building directly atop earlier, government-approved blocks. Orphaning the block would mean the renegade miner’s reward, their coinbase transaction, would also be orphaned, resulting in the miner having nothing to exhibit for their efforts.
The subsequent events are unclear to me, but none of the potential outcomes are favorable. We would face a fork of some sort. The new fork could employ a different algorithm, rendering all existing ASICs incompatible with the new chain. Alternatively, the fork might retain the existing algorithm but invalidate blocks originating from identified malefactors. Either scenario would result in a government-compliant bitcoin and a non-compliant bitcoin, where the compliant fork would continue operating under the original code.
When I have heard Bitcoiners discuss these circumstances, they typically assert that everyone would abandon “government coin” in favor of “freedom coin”. But would that truly transpire? Perhaps we, the audience of Bitcoin Magazine, freedom advocates, and cypherpunk enthusiasts, would trade the censored fork bitcoin for the new freedom variant. However, I have my doubts that BlackRock, Coinbase, Fidelity, and the rest of Wall Street would act similarly. Therefore, the relative economic significance of these two forks, particularly five to ten years down the line, is far from evident. Even if a non-compliant bitcoin fork were to endure and maintain a substantial portion of its economic value, it would be economically and philosophically diminished.
Now consider the same assault scenario but with a well-distributed hashrate. Suppose U.S. miners constitute only 25% of the hashrate. Imagine the U.S. government compelling miners to blacklist addresses, and even worse, orphan any new blocks containing transactions with those blacklisted addresses. This situation is still detrimental. Yet, the 75% of miners outside the jurisdiction of U.S. law would continue to incorporate non-compliant transactions, ensuring the heaviest chain still includes non-compliant blocks. In this distributed-mining scenario, if a fork occurs, it would be the government-compliant bitcoin that would have to fork away and relinquish proof of work for social consensus.
This remains a grim scenario. Custodial services within the U.S. may be coerced into supporting the new compliant bitcoin, posing an economic risk, at least temporarily, to authentic bitcoin. However, if the mining network survives outside the U.S. and holds the majority of hashrate, this scenario seems more akin to the U.S. opting out of bitcoin rather than co-opting it as it could with hashrate supremacy.
What Led to Bitcoin Mining Being Situated in Large U.S. Data Centers?
The progression of Bitcoin mining serves as a case study in economies of scale.
Let’s revisit the origins. The functions we associate with miners — aggregating transactions into blocks, performing proof of work, and disseminating their blocks to the network — were all integral to Satoshi’s descriptions of node operations. There weren’t distinct “miners”; every node could initiate mining with a simple click. Thus, during those initial days, mining was as decentralized as the nodes themselves.
However, CPU mining was swiftly overtaken by mining through graphics cards and FPGAs, and subsequently from 2013 onwards, by ASICs. Mining remained a marginal option on nodes for several years, until in 2016 Bitcoin Core decisively removed the pretense and eliminated it entirely in version 0.13.0 of the software. Once mining gained independence from node operation, utilizing specialized equipment and expertise, it began to scale. This transition was entirely foreseeable.
In The Wealth of Nations, Adam Smith illustrates a pin factory that employs merely 10 individuals, producing 48,000 pins every day, where each worker, independently, could produce a maximum of 1 pin daily. By concentrating on a specific step within the pin-making process, enhancing tools for each individual task, and coordinating their efforts sequentially, the workers created significantly more pins with the same level of manpower. One way to consider this is that the expense of increasing output by one pin is trivial for a factory already manufacturing 48,000, having already invested in the machinery and expertise; it would only necessitate a minor increment of labor and materials. However, for an individual producing a single pin each day, the marginal cost of adding one pin to production doubles.
Mining, once liberated from the CPU, featured numerous aspects that facilitated economies of scale similar to pin production in a factory. ASICs function as specialized tooling, akin to pin-making machines. Data centers also are tailored for the unique power density and cooling demands of those ASICs. Similarly, compared to mining in one’s basement, operating in a multi-megawatt commercial facility allows the same fixed expenses to be distributed across many more mining units. Some instances of relatively scale-insensitive costs faced by miners include:
- Power expertise
- Power equipment
- Control systems expertise
- ASIC repair expertise
- Cooling expertise
- Cooling facilities
- Legal expertise
- Finance expertise
In a larger enterprise, not only are fixed costs mitigated by a greater number of revenue-generating machines, but there is also increased negotiating power with suppliers and labor. Transitioning from one’s basement to the nearby commercial park results in a more favorable electricity rate. Advancing from an office park presence to a mega-center allows one to employ power specialists who devise intricate contracts with power suppliers and financially hedge against price fluctuations. Sending a single machine for repair each time it fails incurs a higher cost per repair incident than merely hiring a repair specialist to identify failing ASICs and address them on-site, provided the operational scale is sufficient. Moreover, when negotiating with ASIC manufacturers, pricing correlates with the size of the orders. Major players can negotiate tougher terms, squeezing smaller miners much like Walmart pressured local shops by securing lower prices for their products.
Economies of scale should not astonish anyone as they are applicable to some extent to nearly all manufactured commodities. The advantages of size readily elucidate how mining transitioned from an activity I engaged in with graphics cards in my basement 13 years ago to facilities nearing 1 GW today.
However, this explains why mining has expanded, not why it has consolidated in the U.S. and among large public companies. To comprehend the latter necessitates recognizing two additional factors. The first is another commodity that scales: financing. Large public enterprises can generate funds by diluting their shares or issuing bonds. Neither of these fundraising approaches is available to a small-scale miner. Indeed, they can secure loans, but not under the same terms as a larger corporation, and the U.S. possesses the most extensive capital markets globally. Secondly, the U.S. operates under a “rule of law,” a relatively stable legal framework, which diminishes the risk that, for example, the state would confiscate a mining operation or that regulators would capriciously suspend operations.
The other aspect that attracted mining to the U.S. in recent years was the availability of power infrastructure. Following China’s ban on Bitcoin mining, it became profitable to mine almost anywhere globally with virtually any ASIC. However, the U.S. had an accessible power infrastructure, much of it in the rust belt, abandoned as U.S. manufacturing moved to China. The U.S. also had abundant energy resources in West Texas, with stranded wind and solar power incentivized by subsidies yet insufficiently connected to East Texas and the rest of the country. In the aftermath of the China Ban, miners swiftly occupied the underused rust-belt infrastructure and exploited the plentiful energy and inexpensive land to establish data centers in West Texas.
The capacity to procure and allocate substantial amounts of funding presents a remarkable advantage, compounding with others, given the fixed, global reward of Bitcoin mining. With ample financing from the markets, the largest public Bitcoin miners could secure the most recent, efficient, and powerful ASICs while negotiating the best electricity contracts and employing top-tier firmware and software experts, among others. This not only put smaller miners at a disadvantage, but also enabled large miners to significantly boost global hashrate, escalating difficulty. When the price of Bitcoin dropped, with a debt-financed ASIC fleet already in operation, profit margins for miners lacking the benefits of scale shrank to nearly nothing. Even a public miner in bankruptcy could continue running their extensive array of machines during restructuring, eliminating smaller competitors while navigating the legal framework.
Thus did mining evolve from a hobbyist scale to gigawatt scale, and thus did it establish itself in America. Mining is a fiercely competitive commodity industry, and the efficiencies provided by scale proved to be pivotal, especially when funded through debt and equity dilution.
Why Mining Will Be Distributed and Small-Scale Once Again
Just as there are economies of scale, there also exist diseconomies of scale, where the costs of unit production actually rise with size after a certain threshold. For instance, it is clear why there isn’t merely one massive food factory that prepares meals for everyone worldwide. Yes, there are efficiencies in food production from factories — observe the average farm size over the past century — but limits also exist. Fresh ingredients need to be transported to a factory, and the final product must then be delivered to consumers. Both the inputs and outputs of a food factory are perishable and heavy. The expenses associated with shipping to and from a single factory would be astronomical, and quality would degrade compared to more localized markets with fresher produce. Similar reasons clarify why sawmills and paper mills are situated near forests, and why bottling plants are located close to fresh water.
However, shipping bitcoin incurs no costs: it is a straightforward task of making a ledger entry on the Bitcoin blockchain itself, which takes mere seconds. While I enjoy boasting about mining our artisanal Portland bitcoin, there are actually no geographic variants of bitcoin that differ based on the location of production. All bitcoin is qualitatively the same. This further emphasizes that global bitcoin production should centralize in the single, optimal location for bitcoin creation.
There’s just one complication with centralizing all mining within a solitary facility: Bitcoin mining is energy-intensive. In fact, it currently consumes over 1% of the world’s electricity. Electricity constitutes the primary operational expense of bitcoin mining, frequently accounting for 80% of operating costs. And unlike bitcoin, electricity is not easily transportable. Not at all. In reality, electricity resembles food that spoils instantly and necessitates costly, specialized infrastructure to transport. For electricity, that infrastructure comprises wires, transformers, substations, and various components of an electrical grid.
Transporting electricity constitutes a significant portion of the total cost of electricity. What we refer to as “generation” often accounts for a minority of the overall expense of electricity, which also encompasses “transmission and distribution” fees. And while the generation costs continue to diminish with advancements
in technology and manufacturing productivity for solar panels, grid investments are just becoming more expensive. Therefore, it doesn’t make sense to transport electricity globally to a singular bitcoin facility. Rather, bitcoin facilities should be situated at the generation locations where they can completely evade transmission and distribution expenses, and subsequently transfer the bitcoin from those locations at no cost. This is already occurring, in fact. It is referred to as placing your Bitcoin mine “behind the meter”.
Mining firms will emphasize their distinctions: firmware, pools, cooling setups, financial acumen, power knowledge, management teams. However, at the essence of what they do, there is minimal differentiation among various mining firms: The output is the same, it incurs no shipping costs, and they utilize the same machines (ASICs) to transform electricity into bitcoin. Variations in electricity expenses largely dictate which miners will thrive and which will falter. During a protracted phase of price stagnation, or even a gradual increase, only those companies with access to the least expensive electricity will remain operational.
The principal argument, therefore, for a worldwide allocation of miners in the future is as follows. Firstly, Bitcoin mining, by its very nature, is compelled towards the most affordable energy on the planet. Secondly, inexpensive energy is distributed globally, including “behind the meter”. Hence, thirdly, mining will also be distributed geographically and behind the meter.
For the sake of discussion, envision that Donald Trump’s wish is fulfilled and all mining takes place in the U.S., with mining in balance, i.e., mining margins are extremely narrow. If someone discovers cheaper power elsewhere in the world compared to the average U.S. miner’s, and deploys ASICs there, the hashrate will expand and some U.S. miners (those with the highest expenses) will cease operations. This cycle will continue until mining is exclusively conducted in locations with the cheapest energy on the globe.
Low-cost energy manifests in various forms: gas in the Middle East and Russia; hydro projects in Kenya and Paraguay; solar in Australia, Morocco, and Texas. The reason energy is distributed is that nature has arranged it that way. Rain and elevation changes (i.e., rivers) exist everywhere. Fossil fuel reserves can be found globally. Wind is a constant presence. The sun shines almost universally.
In fact, the worldwide distribution of energy is somewhat secured by the solar trajectory around the globe. As the sun shines most brightly, its energy is inevitably wasted by solar-powered systems, as power infrastructures are never equipped for peak generation. I anticipate that in the future, a significant portion of the hashrate will trail the solar path, with machines utilizing excess solar either overclocking during that time or, if they are older and unprofitable otherwise, activating only for that brief period when the system produces surplus electricity beyond what the grid requires.
The primary argument can be slightly adapted to arrive at other conclusions regarding the future of mining. I also believe, for instance, that there is plentiful cheap power on a smaller scale, and a limited amount of inexpensive power at a genuinely massive scale (100 MW+). It follows that, assuming Bitcoin mining continues to expand, small-scale mining will reemerge and the trend toward megamines will reverse as substantial sources of cheap power diminish.
To understand why cheap power predominantly exists at a smaller scale, we could analyze specific instances. For example, we might explore why flare-gas waste occurs in a distributed small-scale manner, and why solar inverters are often undersized, leading to clipped power outputs across the system. However, I prefer to contemplate the larger principle. Where we see cheap power at scale, it often represents a significant error. For instance, the error may involve constructing a dam or nuclear power station that was never truly necessary. Major mistakes are restricted in number: They are costly! There is a ceiling to fiat foolishness.
Smaller-scale mismatches of supply and demand will become more frequent, all else being equal. If gas production at an oil well is sufficiently substantial, for instance, it will be sensible to construct a pipeline for transport; if it’s relatively minor, a pipeline will not be practical and the gas will be left stranded. The same holds true for landfills. The largest landfills possess generators and are grid-connected, but smaller landfills frequently struggle to even collect their methane, much less generate electricity with it and channel that electricity to the grid. The same applies to dairy farms.
Moreover, bitcoin is not the sole form of energy-intensive computation. If substantial quantities of low-cost energy are available, other types of computation will take up residence there, and being less sensitive to electricity prices, they will outbid bitcoin miners. Those alternative forms, at least currently, do not scale down as well as bitcoin. It follows that the era of mining on supercheap, large-scale power is limited. Conversely, if you are mining bitcoin by mitigating flare gas on a desolate, wind-swept oil site away from a pipeline, there is virtually no chance anyone will outbid you to conduct AI inference in your area. The same applies if you are mining on overprovisioned home solar energy. Small-scale energy waste is far less attractive to competitors but is useful for Bitcoin miners. Mining can scale down sufficiently to penetrate these pockets of energy, whereas other kinds of energy consumers cannot.
Another iteration of the argument above hinges on the distributed demand for waste heat. All electrical energy entering a bitcoin miner is retained and exits the miner as low-grade heat. With this waste energy, miners are warming greenhouses, communities, and bathhouses. However, heating requirements can usually be satisfied with a small deployment of machines. An ASIC or two can heat a household or a swimming pool. Nonetheless, utilizing waste heat to replace electrical heating enhances the overall economics of mining. All things being equal, a miner selling their heat will be more lucrative than a miner who does not sell their heat. Hence, here is another rationale that mining will be globally distributed and on a smaller scale: The demand for heat is distributed globally — although more so in the far north and south — and at a very limited scale.
As I’ve indicated, I contend that Bitcoin mining will be directed to the world’s least expensive energy. However, this is the trend only if the bitcoin price increases slowly. In a vigorous bull market — and we have witnessed several — Bitcoin miners will utilize any available energy, wherever they can connect machines. If bitcoin’s price skyrockets to $500,000, all my projections will be rendered void. Nonetheless, even in this bullish scenario, mining becomes globally distributed, this time not because the cheapest power is spread out but because available power is spread out. Bitcoin at $500,000 suggests all ASICs are profitable at any power source, and the U.S. alone lacks the infrastructure to manage that level of demand shock, even if it intended to. Therefore, bitcoin will be distributed regardless.
It is also important to note that high-margin periods are fleeting, as ASIC production will invariably catch up, in the pursuit of profit, causing margins to decline again. Consequently, in the long run, the distribution of Bitcoin miners will continue to be influenced by the allocation of the world’s cheapest energy.
For my arguments to hold, the diseconomies of scale must surpass the economies of scale outlined prior. To ascertain the balance between these two requires nothing less than an in-depth analysis of the spreadsheets of each variety of mining business, which would be inappropriate here.
In summary, I believe that if the disparity in electricity costs is significant enough, it will outweigh all other factors. However, I cannot claim to have provided anything resembling conclusive proof here. These are the broad outlines; the nuanced details remain an exercise for the reader.
Geopolitics
Up to now, I’ve pondered over miner incentives without considering the nation-states themselves. It’s evident that while some nations are acquiring bitcoin, others are extracting bitcoin using their energy assets. Nation-states possess motivations separate from anything that Satoshi envisioned. For example, Iran might engage in bitcoin mining to capitalize on its oil, as sanctions render selling it on the open market infeasible or, at best, costly. Similarly, Russia may mine for akin motives. Such state actors could “mine at a loss” compared to a miner covering their own electricity expenses, since the nation-state’s energy costs are subsidized by taxpayers. Their extensive mining operations could, in turn, diminish profitability for others and drive marginally profitable miners out of operation.
I do not perceive nation-state mining as ultimately centralizing hashpower, though. Presently, mining in Russia and Iran is indeed beneficial for bitcoin, as it counters the progress of mining by U.S. public corporations, which overshadow them in size. Furthermore, if a particular nation-state starts yielding an unequal portion of the hashrate, while bitcoin remains a significant component of the global economy, I anticipate that other nation-states with interests in bitcoin’s prosperity — or even substantial bitcoin holders — might also commence mining at a loss to maintain mining distribution.
The game theory involved is not straightforward. Instead of a rivalry to dominate, bitcoin represents a scenario in which all participants prosper when no single entity holds power, while everyone suffers when anyone achieves dominance. For nearly all other technologies or armament systems worldwide, the optimal strategy is to attain global supremacy. Thus, we observe a race for hegemony in battery technologies, chip production, drones, AI, and beyond. This phenomenon is referred to as the “Thucydides trap” in international relations since it encourages a preemptive strike against an ascending adversary: The prize for finishing first is enormous, while the cost of finishing second is unfathomable.
Conversely, if you dominate Bitcoin mining, that is detrimental for Bitcoin mining, and thus unfavorable for bitcoin and, consequently, for you. As Bitcoin mining becomes concentrated within a single nation, all participants recognize the threat of an assault on the neutrality of bitcoin, which is central to its value proposition. For instance, Russia might hold bitcoin to prevent the U.S. from freezing its reserves, as the U.S. did with Russia’s fiat assets following their invasion of Ukraine. Yet, if mining is centralized in the U.S., Russia might distrust that its addresses wouldn’t be blacklisted by the U.S. Treasury. Thus, Russia would sell its bitcoin for another asset if it perceived this looming threat. Miners in the U.S. would experience an increase in their share of block rewards as they achieved supremacy over other miners, but the worth of those block rewards would diminish as the price of bitcoin itself dropped. Other circumstances being equal, then, miners in the U.S. would not wish for Russians to cease mining and offload their bitcoin. U.S. miners should refrain from wanting to “win”, at least not in this manner. If bitcoin constitutes a substantial enough part of the U.S. economy, the U.S. should refrain from wanting its miners to win. Instead, if any nation nears dominance, we can anticipate those deeply invested in bitcoin, including nation-states, will mine adequately to safeguard their own investments.
Bitcoin enthusiasts should aspire for the USA to mine sufficient bitcoin that no nation, including itself, controls a majority. That’s a dismal slogan for a campaign rally, and it doesn’t spark the imagination like “hash wars”. However, as a Bitcoiner, it is the sole pragmatic preference one should maintain.
Disclaimer: The views expressed are solely those of the writer and do not necessarily represent the positions of BTC Inc or Bitcoin Magazine.
