[Note: the following article was reposted on Business Insider, a PDF version is available]

One of the contentious areas of writing about Bitcoin data and emerging markets, is discussing what conclusions and interpretations (if any) can be drawn from say, transactional volume.

Let us put that aside for a moment and consider ways to estimate real commercial volume. Are there any other ways to do so besides a full traffic analysis?

Sell side pressure

On any given day there are at least three entities that continuously sell bitcoins onto the market: merchants (and merchant processors), miners and mining manufacturers (who are sometimes paid in bitcoin).

As I noted in my previous article, last month BitPay announced that it was processing about $1 million in daily payments. It is unclear what amount of bitcoins that constitutes, depending on the time frame and therefore price levels (early December or the month of May) it could represent 1,000-2,000 tokens per day.

Let us assume that the other merchant processors such as Coinbase and BIPS are also processing a similar amount. And that altogether between 5,000-10,000 bitcoins per day are collectively being spent on commercial activities through these processors.

This puts pressure on the sell side of the price equation. That is to say, to minimize exposure to volatility, nearly all merchants elect to immediately convert bitcoins into fiat and those bitcoins are sold onto the market (both Ben Edelman and David Evans have written on this before).

Similarly, because miners have to pay real costs – capital and operating costs – they too sell their mining rewards on the market: around 3,600 each day (because again, MV=MC).

It is unclear how much mining manufacturers have to sell each day to fund their own developmental and logistical operations, but for the sake of simplicity and roundedness, let us say 1,400 bitcoins (it could also be as little as zero).

Thus altogether, there is a regular 10,000 – 15,000 bitcoins representing commerce that are sold daily on the market today. It also bears mentioning that, although technically the miners receive money, virtually all of it is spent towards utility (electricity) and hardware, not on the bitcoin ecosystem itself. While it is unclear how much other positive-sum value exchange is taking place (such as remittances, houses or cars) we can see that the transactional volume of potential commerce has remained flat during the past six months:

Is there a chart that shows this amount of transactions?

In my last article, I mentioned Total Volume Output – the total value of all transaction outputs per day – yet this includes coins which were returned to the sender as “change” and thus the real number trying to be measured is substantially less. And taken to its maximum readings, roughly 1,000,000 bitcoin outputs (UTXOs) are used each day.

If only 10,000 – 15,000 bitcoins are being used in real commercial activities (instead of merely zero-sum activities like gambling, mixing of coins or cybercrime), then the perceived Total Volume Output is potentially two orders in magnitude larger than the real economy.

What is the real economy? While the debate over what percentage of bitcoins are being spent in positive-sum activities, between October 15 and December 18 of last year, 41,928 bitcoins were sent to addresses controlled by Cryptolocker (a type of malware) – this is not real economic growth, in fact it is negative-sum. And because it signaled to the market that it was a successful way of generating (stealing) wealth, there are numerous copycats using similar methods (including CryptoDefense and Cryptolocker 2.0).

The cost of information security

For the moment, let us ignore the buy side of the equation, that in order to keep the same price level, at least 10,000 – 15,000 bitcoin are being acquired by other parties each day (primarily high-net worth individuals and institutions through OTC brokers).

What this actual activity translates into is the following:

Miners are the labor force that secures and processes transactions. And because this labor force has real depreciating capital costs and operating expenses, in theory, the cost for their services amounts to roughly $2 million per day (3,600 bitcoins X $600 per bitcoin).

In practice however, most miners are operating at losses. In fact, the network is vastly oversecured by miners operating at losses probably by a factor of 2-5x (described in Estimated costs). For instance, according to a recent report from the National Science Foundation (NSF), a now-banned researcher used, “about $150,000 worth of NSF-supported computer use at the two universities to generate bitcoins worth about $8,000 to $10,000.” Or in other words, the researcher externalized the real costs of mining (energy and capital deprecation) onto another party (the NSF and therefore taxpayers). This is inefficient, yet there are many cases of such activity taking place each day.

Thus while the Bitcoin ‘trust fund’ (a more accurate description for the network which divvies out a finite amount of block rewards) pays out security of $2 million each day, the labor force is providing significantly more security than they are being paid, probably closer to $6 – $10 million if not more (Hass McCook has additional estimates).

Simultaneously, they are providing these services for commercial activity that ranges from as little as 5,000 bitcoins to perhaps as high as 15,000 bitcoins. Or $30 million to $90 million respectively in today’s prices.

For comparison, MasterCard spent $299 million on their capital expenditures in 2013. As part of these expenses, it builds data centers similar to the “fortresses” (with moats) that Visa has also built. In 2013, MasterCard and Visa processed a combined $7.4 trillion in purchases.  Together with American Express and Discover, these four companies generated $61.3 billion in revenue during the same period.

While this is not an entirely apples-to-apples comparison, what this means is that the Bitcoin network is enormously oversecured compared with other transactional platforms. The reason this is, is because it is decentralized which creates overhead (since all the nodes have to process and verify the transactions). Yet, as shown with GHash.io over the past month, the network is qualitatively insecure due to economies of scale. That is to say, so as long as the proof-of-work mechanism can be economically scaled, this leads towards centralization. No amount of white papers or tweets will change that.

If the labor force of bitcoin is spending $10 million on protecting the network yet real commerce is only $30 million, this would be equivalent to a mall issuing 1 out of 3 customers a personal security detail to go shopping. Or in other words it is, arguably, quantitatively oversecure (it is not qualitatively trustless as shown by GHash.io and previously, Deepbit). Perhaps this mix will change over time. However one thing to consider is that some bullish advocates contend that the Bitcoin network will one day supplant and compete head on with Paypal and even Visa. In order to do so, the Bitcoin labor force are still (assumedly) being paid a fixed income to provide the same services. Thus perhaps in the future, the opposite will occur – the network could become undersecure due to disproportional rewards.

I spoke with Greg Simon, co-founder of Cryptowerks who worked as head of International Equity Sales for JP Morgan in Japan. According to him,

Cryptoledger miners are Japanese banks. They are producing an oversupply of crypto trust relative to an under supply of borrowers of that trust. Their only solution, producing an ever increasing supply of crypto trust, is making the problem worse, not better. It is the equivalent of central bank QE [quantitative easing], or pushing on as string. Just as an oversupply of central bank produced money causes the value of each unit of money to decline, so does an oversupply of crypto ledger miner produced crypto trust cause the value of each unit crypto trust, which we can measure in units of ghash, to decline. The problem is not the aggregate supply of crypto trust. The problem is aggregate demand for crypto trust. Until demand for crypto trust improves, either from monetary or non-monetary borrowers, we can expect the same fate for crypto trust in the crypto economy as we are seeing for fiat money in the legacy central bank fiat economy.

Another way to visualize this phenomenon is the chart below:

John Ratcliff recently published an explanation about zombie bitcoins (coins, or rather UTXOs, that have not been active in more than 18 months) which is where the chart above comes from. Each color band represents the last time a private key corresponding to these UTXOs was used.

Thus, one take-away from this chart is that liquidity – as shown by the One Day, One Week and perhaps One Month bands – represents between 100,000 to 2,000,000 bitcoins. What is the actual number? Without a full traffic analysis we probably will never know.

But we can tell from spikes that the largest movements take place during volatile time periods, specifically during price run-ups. So, for instance, in the spring of 2013 there was enormous Western media attention and a subsequent boom that peaked in mid-April (when Mt. Gox had to temporarily shut down). Similarly, in November and early December corresponds with additional global media coverage and Chinese adopters coming online – with prices peaking on December 4^th. Or in other words, transactional volume rises and falls with price levels – that the bulk of on-chain activity corresponds primarily to day trading and speculation. This, despite the fact that Ratcliff notes, that prices during this 18 month time span increased 4,000%.

That is to say, even though there are more than 100,000 merchants that accept bitcoin and even though token valuation has risen logarithmically, UTXO holders as a whole prefer speculating over conducting actual commercial activity. What could change this behavior?

Maybe nothing will because Bitcoin is a recreation of a medieval agrarian economy; few people spend, in part because the network codifies what is essentially negative time value of money.

Money and credit

There is an endless stream of papers and books on the topic of what constitutes and attributes of money. Arguably one of the most thorough explanations of what money is and how it arose is, The Ascent of Money by Niall Ferguson which was later turned into a good PBS series.

Despite what some Bitcoin advocates claim, gold itself was not used on a large scale since time immemorial. In practice, there were numerous types of physical assets ranging from metals to stones. England even used a system of money known as tally sticks for several hundred years. And the reality is that prior to the birth of civilizations, many tribes and villages operated with barter and gift systems with themselves and one another (some never even created something akin to “money”).

As noted by Ferguson, up until the Renaissance, there were no real financial instruments or professionalized banking or hedging methods in the West. Bonds, joint-stock corporations and insurance companies evolved throughout time (all post-Fibonacci). And consequently, this is reflected in the dearth of economic output at the time. That without a way to expand credit – to create loans to start businesses – the pie cannot be enlarged. In his words, “Credit and debt, in short are among the essential building blocks of economic development, as vital to creating the wealth of nations as mining, manufacturing or mobile technology.” In contrast, poverty (subsistence) more often, “has more to do with a lack of financial institutions, with the absence of banks, not their presence.”

Thus, I would argue that ultimately Coinbase could turn into a fully-fledged bank, providing interest to bitcoin holders to be able to loan out bitcoins (much like BTCJam does). And they could do this through a fractional reserve process. In fact, Huobi’s new Hong Kong branch (BitVC) is heading in that direction; users can lend funds to Huobi for interest and Huobi will then lend it out to users to trade on margin. Contrary to what many Bitcoin adopters contend, fractional reserve banking itself is not inherently a bad thing but that is a topic for another article.

However, for the time being, Coinbase and others – while on-ramping a lot of new users and providing utility through ease-of-use functionality – have a long way to go before this occurs. For instance, this past week Coinbase announced instant buyback (once you spend bitcoins, you can buy more). However, the reddit comment below sums up actually what happens:

Again, there is a difference between the economy Bob wants to have versus the economy Bob currently has. Today Bitcoin, as I have argued, is at most an emerging market akin to a pre-industrialized agrarian economy with enormous frictions. Internally it is an inflexible command economy that outsources and arbitrarily rations its scarce resources (block rewards) irrespective of economic conditions (e.g., Bob, the miner, is rewarded whether or not he processes transactions).   Front loading rewards the first four years without processing any transactions is an unsustainable activity. In fact, as Jonathan Levin, co-founder of Coinometrics, notes in his upcoming paper, Creating a decentralised payment network, he found that “[i]n total over the network history there have been 84,469 blocks with no transactions.” Yet because there is no one at the helm, no entrepreneur to rationally allocate block rewards or market value for those rewards the first year, ultimately 4.2 million bitcoins were given out for naught.

Many adopters note that this was done to help bootstrap the economy and that the initial distribution of bitcoins through the block reward is purportedly not how bitcoin will operate in the long run. And that at some point Bitcoin’s internal economy will somehow be incentivized by transaction fees only – or at least that is theoretical transition (see Reducing and removing block rewards). But the fact that miners were rewarded irrespective and arbitrarily of their actual work is very similar to how top-down command economies work rationing wages. This is a topic that will likely be debated over the coming years.

For additional perspective I spoke with Martin Harrigan, a software developer and founder of Quantabytes, a cryptocurrency analytics start-up. In his view:

The initial distribution of bitcoins is a one-time process that is distorting our understanding of the Bitcoin economy. I think that the peaks in transaction volume during the price run-ups are a form of secondary distribution: early adopters are distributing their bitcoins, for profit, to new users. This is a vital part of Bitcoin’s distribution process and may continue for as long as there are periods of significant price increase. It may be that institutional investors will take-over a significant portion of this process for several years. Then, at some point, when the technology, infrastructure, regulatory frameworks, and our understanding of cryptocurrencies has matured, the price will stabilise and Bitcoin will return to individual users as a stable transactional currency in the traditional sense.

Of course, I’m speculating wildly here. My point is that we don’t have a good null model. We’re not seeing “hockey stick growth” but maybe that’s okay. Many start-ups need this type of growth to survive — I don’t think Bitcoin does. During the Bitcoin crash of 2011 the price dropped 93% and didn’t recover until 2013. The difficulty also dropped and remained stagnant for a year and half. Although I can’t quantify it, the “mood” on the Bitcointalk forums was grim. The equivalent event would have been fatal to most start-ups.

Perhaps at Harrigan noted, this will change in the future. And perhaps those frictions are still lower than the cost of doing business in certain regions (like the Philippines). And this is not to single-out Coinbase. I still think they are one of the best companies in this space, I even consider them one of the most promising in part because they are trying to create value (and have). But that does not mean this particular service is frictionless. For instance, David Evans recently delved into the specific transaction costs of various platforms and explained how the actual roundtrip cost of Coinbase is 2% + $.15, not the 1 percent that is frequently cited (Evans also had another good explanation of how Coinbase and merchants like Overstock work).

Building a continuous series of bubbles or building utility

One of the common refrains about altcoins and appcoins in general is that none of the underlying systems are able to give out real equity and thus cannot have P/E expectations – neither does bitcoin, nor will it. This is a bug and it is why I argue that using the “TCP/IP” analogy is probably incorrect.  The internet is an amalgam of private-public intranets cobbled together and cost real capital to build.  It was not built with magic; real incentives had to be provided to build it.  Imagine if those incentives decreased 50% every 4 years?  That’s Bitcoin’s internal economy. The Bitcoin network cannot operate without bitcoins – the app or currency or commodity (choose your definition). The two are united together. Yet TCP/IP, the protocol, can still work even if substantial portions of the network fail; it is not tied to a specific set of hardware or token (TCPIPCoin).

This is not the first time a set of unrealistic expectations have been created in the past 10 years.

So what kind of bubble is bitcoin then? Some claim that it “crashes upwards” which makes no rational sense at all. Bitcoin (the token) is not immune to the laws of economics. Perhaps as illustrated above (source), the current investment cycle in Bitcoin is more akin to Cleantech circa 2005? What this means is that, as noted in my previous article, even though many of the startups are clever, they may lack sustainable business models. Once this froth is removed, the businesses that survive will likely be those that are actually creating real pain killers (utility) to real needs; perhaps reusing the infrastructure of the network (like merged mining proposed by Blockstream or that of Cryptowerks) to process contracts and titles. Again, all of this is speculative, yet it warrants attention because Cleantech also had a similar dedicated ideological group of early adopters that created economic activity (though, not much growth yet) and wanted to change the world. And despite their best efforts it popped.

In conclusion, expanding credit alone is not the answer to Bitcoin’s stagnant economy. For instance, China’s money supply grew leaps and bounds since November 2008 (when it implemented a series of stimulus packages). It also signed bilateral currency agreements with new countries every year which led many outside commentators to erroneously conclude that this somehow leads to mass adoption of the RMB. Yet the stark reality is that the RMB only accounts for 1.4% of global payments compared with the dollar at 42.5%. This is unlikely to change either. However, even in its current doldrums, the Chinese economy still produces real goods and services to the tune of trillions of dollars per annum. Obviously it is unfair to compare Bitcoin, a five-and-a-half-year old “startup” to China. Yet the emerging market aspect, the reuse of capital stock, the implementation of new financial instruments, the training of unskilled laborers and ultimately the creation of needed utility to outside parties can be viewed as facsimiles to learn and grow from.

As noted by John Kenneth Galbraith in the last article, there is only so much capital that can be extracted from the “fleece-me” crowds of reddit. Significantly more capital is needed to scale operations to enterprise-level reliability. While some advocates believe eschewing the ancien regime of venture funds and private equity is the way to move forward, this is short-sighted.

Below is a list of companies I think have potential to create value in this digital ecosystem – embryonic solutions to this quagmire (I hold no equity in them):

• API: Chain, Blockr, HelloBlock, BlockCypher, HiBitcoin
• Analytics: Coinalytics, Coinometrics, Blocktrail, Quantabytes
• KYC: CoinTrust, Block Score, Coin Validation
• Decentralized cloud: MaidSafe, StackMonkey, decloud, Bitcloud, StorJ
• Lending: BTCJam, Bitreserve, Bitfinex
• Peg: Netagio, Ripple Singapore, Digital Tangible Trust

I have a new article that went up over at Let’s Talk Bitcoin yesterday called, Can Bitcoin change from a bubble economy into a growth economy? (pdf).  It is essentially part 2 to the previous piece and put together based on feedback I received this past week.

I should mention that while I did use an equation in the middle of the article to describe prices, I do not think the regression itself is valid.  As my friend RD mentioned, the error term does not seem random. There is a deterministic trend that would have to be filtered out first.  It is the same problem with modeling long term GDP growth.  After all, why would we need all these economic models if we can simply draw a straight line and predict GDP using high school math?  Many things follow an exponential growth curve, that is nothing new. The exciting thing is to forecast it in the short term, which this method is extremely bad at (the same criticism can be lobbed at models like Elliot wave theory).

Over the last 12 hours I have received some criticism about one particular point: zero-sum games and gambling.

In one email exchange, Bob disagreed that gambling was zero-sum, stating:

Gambling in and of itself does not create productivity, but the businesses that surround gambling certainly can. See: casinos, bitcoin mixing services.

Also, if the house makes a bunch of money, and the owning entrepreneur uses those funds to start another, productive business, then in some sense the gambling has facilitated economic development by liberating wealth from unproductive suckers who participate in online gambling to a highly productive entrepreneur. This was, of course, exactly the case with the most popular bitcoin gambling website.

The issue here is a measurable one.  A zero-sum game is one in which wealth is merely redistributed and not grown.  What Bob described above is economic activity but not economic growth.  Gambling is zero-sum game as is speculating on stocks or cryptocoins, no new utility itself is created.  Tokens are simply being moved from person to person. Eventually many people are left with assets that they cannot sell because all the demand has been fulfilled, and at that point the price may actually crash.  In other words, to make money in a zero-sum game, it is only because others have lost an equal amount.  In fact, in many cases, value diminishes because of interchange fees or in the case of mixing services, transaction fees.

This touches on an economic principle of opportunity costs (the “seen” and “unseen”) — the traditional example used is Alice throwing a brick through a shop keepers window.  While the seen result is a repairman being hired to fix the window, thus spurring economic activity, this does not actually create economic growth because the shop keeper must now forgo certain opportunities to spend repairing existing physical stock.

Note: Gambling has the name “math tax” because it is a tax on people who are not good with statistics (49.5% odds means in the long-run, you will always lose to the house).  This is derived from Ambrose Bierce’s quote, “Lottery: A tax on people who are bad at math.”

After a quick Google scholar search, I think there is more concise explanation of this phenomenon in Gambling and speculation, by Borna & Lowry:

Unproductive nature of gambling

For players, gambling, at best, is a zero-sum game, i.e., the aggregate wealth of the players will not be altered due to a gambling activity. The losses of one party are precisely equal to the gains of the other participants. Of course, if the gambling activity were taxed by the government, or there were other ‘leakages,’ then the expected value of winning would be negative, i.e., the aggregate wealth of the players after the play would not be equal to their original wealth.

Although gambling is a sterile transfer of money or goods among individuals creating no new money or goods, it nevertheless consumes the players’ time and resources and may subtract from the national income. From a macro-economic point of view, the aggregate wealth of the players will change, in the long run, due to the fact that the transfer of wealth is usually among unequal productive sources. It may be argued that the productivity lost due to a transfer of money from one player will be offset by an increase in the productivity of the other player. This assumption is true only if both the winners’ and losers’ production schedules were assumed to be identical and linear.

One last note: there are a number of people I would like to thank for their comments included in this article; this does not mean they agree with or endorse my view.  This includes: Dave Babbitt, RD, Mark DeWeaver, Dave Hudson, JL, Taariq Lewis, CK, Petri Kajander and Chris Turlica.

The past year has seen what some the “Cambrian explosion” in altcoins.  Most are complete pump-and-dump scams.  Others experimentally toy with a couple variables and a remote few actually bring something genuinely new to the scene (like proof-of-transaction in Fluttercoin).

Wouldn’t it be neat to see a list of all the known altcoins and a little bit about what they were before they became ledger rot?

For the last couple of months Ray Dillinger on Bitcoin Talk has been attempting to chronicle this digital death in: Necronomicon thread: Altcoins which are dead.  I wonder who holds the record for creating the most alts?  Is that something you brag about at a bar?

Furthermore, here are some patterns that Ray sees are emerging based on alts that are not (completely obvious) scams:

• Alts have a plethora of different hashing algorithms, but there are really only three divisions that matter between their proof of work.  Either they are CPU dominated, GPU dominated, or they are mined with a particular kind of algorithm-specific ASIC.  ASICs exist for SHA256D coins like bitcoin and for Scrypt coins like Litecoin.  No other algorithm-specific hardware exists, so no other difference in hashing algorithm matters.
• Within each division, coins rapidly reach an equilibrium where whichever hardware generates their proof of work is allocated in direct proportion to the financial reward per minute of the block reward.   Usually this means that halving the block reward of an alt, other things being equal, halves the hashing power securing its block chain.
• Most alts become unstable when 3/4 or more of their coin supply has been mined (second reward halving), even if no other factors have made the cryptocurrency unstable prior to that time.  If the value being secured exceeds the block reward by too great a fraction, 51% attacks and forks are to be expected.
• “Burst mining” or “automatically switching multipools” act like positive feedback amplifiers with a delay loop in mining – they make it VERY difficult for an alt to maintain a ‘stable’ block rate when its hash rate gets too small relative to the pool (GPU, CPU, or specific ASIC type) of hashing power it belongs to.  This is a heavy contributing factor in the 51% attacks and forks mentioned above.
• Several alts switched to proof-of-stake after to this instability led to bad forks or “stuck” block chains where miners essentially had given up due to a burst miner leaving the chain at a ridiculously high difficulty level.
• A shift from proof-of-work to to proof-of-stake, whatever its effect on market value, has usually led to stagnation or decline in the community surrounding the alt.  In turn this usually leads to a collapse in value of the alt, unless heroic measures (such as direct giveaways by stakeholders to non-stakeholders) are used to bring people into the community.

See also: Different proof-of-work mechanisms and several altcoins that have been hit with a 51% attack

A user, Bob, on Bitcoin Talk sent me some question in response to my article:

1) why does one assume transaction fees will substitute decreasing block rewards. are users not equivalent in choosing the operating software in the Bitcoin network why not just charge a mandatory 0.01% charge on all coins that are younger than X blocks and limit free transactions to say 30%. When the thermodynamic limit is achieved economic friction in Bitcoin will be just 0.01% of all energy consumed in the the economy.

2)the thermodynamic limit is not a limit in that there is entropy, the waste heat just becomes an input for a new system, eg. water can be desalinated, cooling as we know it today is a wasteful activity.

3) I also don’t seem to understand why when ASIC chips reach the thermodynamic limit they’ve wouldn’t start decentralizing in location (not ownership). There will always be centralized mining where energy is cheep but there will always be a need for heating at the very least everywhere and that is potentially a free energy input.

The free market is the perfect motivation for innovations not explored in your analysis.  Still you leave me in awe as to how alts are going to evolve as the energy equilibrium evolves.

My responses to Bob are the following:

1) In practice, users of the Bitcoin network do not like including transaction fees and there are endless threads on Bitcoin Talk of people complaining about fees.  In fact, one of the purported — wildly incorrect — selling points with Bitcoin is that it is somehow “free.”  Obviously this is incorrect, utilizing scarce resources is not free.  Someone has to pay.  The people who pay in this case are all bitcoin holders as roughly every 8-10 minutes new bitcoins are minted, diluting the shares of everyone through inflation.

But let us assume that we fast forward 100 years into the future when there are no longer block rewards, that miners will continue providing labor solely for transaction fees.  If these fees are floated and chosen by miners, it is impossible to say a priori what the actual market clearing fee will be.  Will it be 0.1%, 1%, 10%?  Or something in between?

One issue in Bob’s scenario is the “30% free transactions” — this is completely arbitrary.  Miners still have to bear a real cost to transacting and securing the network and only do so today because of block rewards.  If there are no block rewards and they want to continue providing “free” transactions, then they will be doing so out of charity which is not a sustainable business model.

2) The thermodynamic limit is something Andrew Poelstra has written about.  For the purposes of Bob’s specific point, Poelstra’s document can be ignored for the moment because it is still necessary to actually describe what is happening today.  If a bitcoin is worth $1,000, then an economically rational miner will only spend (capital costs + operating costs + taxes, etc.) no more than $1,000 to extract rents on that token.

In practice this is not the case as there are numerous examples of people and companies operating at losses for a variety of reasons, primarily because of price expectations: they believe that the token will eventually appreciate in value and the market value of the token will eventually cover their operating losses.  Thus today, the network is being “oversecured.”

As far as “waste,” that is how Proof-of-work works.  Someone, somewhere has to “burn” (or dissipate) something in order to secure the network.  Irrespective of what part of the supply chain or logistical operations it takes place, market participants are provided signals by token value (e.g. a $1,000 bitcoin) to turn off or on their hashing systems.  It doesn’t matter what the energy source is or how efficient ASICs are, market participants will simply use a calculator to find out if their inputs (capital costs + operating costs, taxes, etc.) allow them to profitably provide their labor.  The same goes for a $1 million bitcoin.

3) Let’s assume that tomorrow several chip manufacturers announced that they were now shipping chips with fabrication node spacing that reaches the Planck limit (see this interesting paper).  That essentially, irrespective of who you bought from, their hardware design was the most maximum efficient chip possible.  We will call this the Alice design.  What would happen in this case is that whoever was able to get a hold of Alice first would profit from it disproportionally at first (as other competing farms were using older less efficient designs).  But over the months, the distribution of Alice became widespread and you could go to a store and buy Alice off the shelf from a neighborhood retailer.

What would happen then is, since everyone is competing with the same hardware, the only variables to profitability would be land costs (plus taxes and compliance costs in your jurisdiction) and operating costs (electricity).  As a consequence, there would be global arbitrage, a dance in which miners would gravitate towards the cheapest region of the globe with favorable tax policies and cheapest electricity prices.

We already observe this happening today, which are discussed in that article.

The benefits that heating may play could be a factor, but if history of cloud computing is any guide, it is relatively unimportant — Google does not put employee housing within the middle of its data center to warm them with a Carnot engine of some kind (yet).  But again, this is unimportant.  All mining facilities, just like any data center, will have a profitability calculator.  Irrespective of how they displace or use the energy they have a bottom line of whether or not they can continue providing the service (via their computational labor) at a profitable rate.

Also, in practice, data centers typically receive subsidies from a variety of sources (like local tax breaks).  Even if you removed all of the subsidies and all geographical regions were “pure free markets” — there are still areas on the planet with better infrastructure, cheaper energy resources, better property rights, etc.  Those are the same locations capital moves to on a yearly basis.