Correlation between mining costs and Bitcoin value and ecological nightmare

[TiagoTiago]

Won't the more direct relation between electricity consumption and profit result in more efficient hardware that use less electricity, as well as an increased stimulus for reaching for sources of electricity cheaper than what big power companies provide?

[PatrickHarnett]

Sorry, I find that funny.  Surplus energy - not really.  Most systems use some kind of fossil fuel and they would prefer to hold it until someone places enough alue on it to consume it.  They are not likely to throw away money on mining.

Oh but surplus energy does exist - I happen to live in a country where about 60% of the electricity is generated by hydropower-plants and they are not turned off at night just because electricity demand drops. Instead they built lots of pumped storage systems to salvage at least some part of the surplus energy to use during daytime.

Going a bit off topic - I too live in a country with greater than 60% hydro electricity, and I am aware of the pumped storage systems around the world too.  However, they cost serious money and they would prefer not to use it if they could avoid it.  Pumped storage is used because it's a lower cost option, not because the electricity is "spare".

[flug]

Won't the more direct relation between electricity consumption and profit result in more efficient hardware that use less electricity, as well as an increased stimulus for reaching for sources of electricity cheaper than what big power companies provide?

More efficient technology will give more hash power, not less power consumption.

Power sources will have a relatively secondary influence. We need to reduce the power by orders of magnitude if Bitcoin becomes a significant %age of the world economy.

[bitcoinBull]

Won't the more direct relation between electricity consumption and profit result in more efficient hardware that use less electricity, as well as an increased stimulus for reaching for sources of electricity cheaper than what big power companies provide?

More efficient technology will give more hash power, not less power consumption.

Power sources will have a relatively secondary influence. We need to reduce the power by orders of magnitude if Bitcoin becomes a significant %age of the world economy.

ASICs/FPGAs are at least an order of magnitude more power efficient (at least 10x more power efficient) than current AMDs.  And AMDs are more power efficient with each generation. 

When the older inefficient chips are replaced by new ones, that will result in more hash power and less power consumption.

[flug]

When the older inefficient chips are replaced by new ones, that will result in more hash power and less power consumption.

I disagree. See rest of thread.

[wareen]

So, before this thread goes to sleep, I'd like to leave a working conclusion that if the bitcoin price increases 200 fold over the next year or so, bitcoin mining will be using in the order of 1GW of energy, or about 1 large power station.

Stated like this it is really misleading - you need way too many assumptions for this conclusion to be of any worth by itself.
Apart from all the other influential factors mentioned in the thread you also ignored the fact, that mining is no rational investment in a market which seems to be consistently growing all the time (your premise).

Example: consider a starting price of 1 BTC = 1 USD and you invest 10.000 USD in a mining rig which gives you 50 BTC per week. With difficulty following price (your conclusion) you'd roughly end up with 650 BTC by the end of the year, worth 130.000 USD. Including electricity costs you'd have a mere 11 fold return on your investment compared to a 200 fold increase if you had bought BTC directly to begin with.

You see that you have to also make very specific assumptions on the expectations of the future price level for predicting the amount of new hashing power entering the market. If for example, miners suspect a pump and dump by big speculators (which would not look too differently to the price explosion you base your argument on), new miners might not enter the market despite its seemingly short-term profitability.

To conclude from my perspective:
Yes, if price shoots up 200 fold within a year, Bitcoin's total energy consumption will be much higher by the end of that year but your conclusion of a 1:1 relationship is ignoring so many factors that it most certainly does not even hit the right order of magnitude.

Also, regarding the usefulness of extreme extrapolation:

[benjamindees]

There is absolutely no way for Bitcoin mining to consume more than even a fraction of the electricity produced in the world today.  We will have assimilated all available GPUs and run out of chip manufacturing lines to build new ASICs before we even come close to that scenario.

[TheGer]

Bedini motors for everyone!

[blogospheroid]

Flug's point is an important one, but nonetheless there are few ways to avoid power consumption, ceteris paribus.

Bitcoin's value is because it is difficult to fake. As more and more wealth gets into the network, every possible exploit will be found. ASICs and FPGAs will drive GPUs out. But if the tech is shared or parallel tech is developed, then power usage will be back to where it used to be. Hash rate will be at a level where ordinary people won't be able to touch it, but power consumption will continue be correlated. Even if something incredibly power efficient like DNA based computing is developed, there will be a large temporary dip, but the power consumption will go back to a high level.

This is because of an exceedingly simple reason - Every automatic way of making money in the world will eventually be tried out. Mining is an automatic way of making money in the bitcoin world.

Money is the fulcrum of the world economy. Billions of people decide what they want to do daily based on money. It is the ultimate coordination mechanism. Many people see this power consumption as wasteful, but I see true agoric commerce as a wonderful new way of coordinating society.

We wish we could trust everyone to the extent that their word is enough, but in reality, we don't.

We need proof of resources, which translates into the world of the internet as Proof of Work.

[mouse]

OP, this has been discussed a few times that I'm aware of, and since I'm new, probably dozens of times.
I tend to agree that this aspect of Bitcoin is a serious flaw.

As money is moved into bitcoin its value will increase.
As the value increases the profits from mining will increase.
More miners will then enter, increasing the difficulty, until the profits are minimal.

Unless I have the relationships wrong, the worlds (environmental) demise is tied to Bitcoins success.

[flug]

To conclude from my perspective:
Yes, if price shoots up 200 fold within a year, Bitcoin's total energy consumption will be much higher by the end of that year but your conclusion of a 1:1 relationship is ignoring so many factors that it most certainly does not even hit the right order of magnitude.

This isn't mindless extrapolation. This is concentrating on the primary market driver.

Cost of mining will tend to gravitate to the rewards of mining.

In a decade's time, when (the few) people who can afford it have invested in expensive ASIC mining rigs and paid them off, and the bitcoin price has stabilized at $10,000 to become a major world currency:

(a) cost of mining will be mostly power costs, not capital costs
(b) investing in bitcoins won't make easy money anymore.

So I think we will approach a 1:1 relationship. At least, it will definitely be in the same order.

I think the calculation you gave is important in understanding the problem, but only looks at at small time window during the initial stages of Bitcoin. In the long term I think the picture will approach a different steady state.

[flug]

Unless I have the relationships wrong, the worlds (environmental) demise is tied to Bitcoins success.

Consider that the block reward halves approximately every four years. Eventually the power consumption (motivated by transaction fees) will settle at a level which perfectly balances the cost and the benefits. Why would one want it to settle at any other level?

Yes, there will be balance. But in 8 years time, the block reward will be 1/4, whereas the bitcoin price will be x1000+ if it becomes a major world currency.

[flug]

In summary.. if bitcoin becomes a major world currency its price will soar -> mining rewards will soar -> amounts spent on mining will soar -> amount of energy consumed will soar -> number of power stations will soar.

I don't think there's any way round this.

Better hardware doesn't solve this.

[ElectricMonk]

Highly competitive markets have near zero profitability... see http://en.wikipedia.org/wiki/Perfect_competition#Profit

For mining to be profitable...

$/฿ > (kWh/฿ x EP)

where

EP  - is $/kWh

Assuming EP remains fairly stable then an increase in Bitcoin value will have a near 1:1 relationship with the energy consumed to mine them.

This is the biggest issue to face bitcoin by a long stretch.

*** Look away now if you're averse to slightly lunatic ideas. ***

Bitcoin value is in proof-of-work. For bitcoin to progress it requires that the energy consumed in generating them is not wasted and that in turn requires that as we consume energy it is measured. Therefore, the future of Bitcoin isn't a new range of Alpaca socks, it's the efficient conversion of mining rigs into useful objects.

*** Really nutty tin-foil hat stuff ahead ***

Imagine a company that provides 'electric heaters' to hospitals/schools on the basis that they connect those 'heaters' to the internet and that the heating company will pay for 50% of their heating bills. Assuming those 'heaters' are as efficient as the thing they're replacing - congratulations! - you've just reduced your mining rig costs by 50% and the school is only paying half of what it used to.

Time for a nice cup of tea. Now where did I put that Bitcoin hashing kettle?

MONK'S FINAL THOUGHT
=============
*Sits on chair and looks out at audience *

If Bitcoin were to become successful then the real achievement would be it's ability to act as a metric of real work done. Those who
find novel ways of converting the hashing energy into something useful will have a huge competitive advantage over those who waste it heating up their garage. That's competition and profit seeking in action. Ultimately the rewards will go to those who do/make something with all this energy and I think that's kinda nice.

Capitalism - I'm lovin' it.

[BubbleBoy]

Using miners as heaters is pretty insightful. Sadly it can't work in real life. At 60 centigrade internal temperature the GPUs fry, and waste heat from coiling them is probably 40-45 degrees at most. That's not really very useful for heating, maybe only if you blow the heated air directly into the rooms. Electric heating is also very expensive, and is often employed in a heat pump configuration that uses electricity to drive a motor (mechanical work), and achieves much better efficiency than simple Joule heating. It's also not needed year-round.

While we fixate on electricity as the main waste for Bitcoin, in reality the capital costs of designing and manufacturing rigs will be quite high. Allot of resources will go into manufacturing, with the associated environment impact. Electronics manufacture requires rare, pure, toxic substances, precision industrial machinery, world wide shipping etc. If Moore's law holds for mining bitcoins, the old hardware might be tossed before it has a chance to equal the environmental footprint of it's production. This is certainly the case with consumer electronics.

[ElectricMonk]

Using miners as heaters is pretty insightful. Sadly it can't work in real life. At 60 centigrade internal temperature the GPUs fry, and waste heat from coiling them is probably 40-45 degrees at most. That's not really very useful for heating, maybe only if you blow the heated air directly into the rooms. Electric heating is also very expensive, and is often employed in a heat pump configuration that uses electricity to drive a motor (mechanical work), and achieves much better efficiency than simple Joule heating. It's also not needed year-round.

I was thinking of ASICs that have a large footprint and a huge heat sink attached to them (i.e. the size of a domestic radiator). Even if they were only 50% as efficient as central heating then the school/hospital would pay < 50% of the electrical costs and you'd have a huge competitive advantage.

I can't see how a heat sunk ASIC could be anything other than near 100% efficient. Where does the energy go if it's not heat? light? sound? potential? chemical potential? Information transmission? Information entropy!! Wow. That'd be something. My guess is that you could convert at least 95% of all that electrical energy into heat. Where else would it go?

While we fixate on electricity as the main waste for Bitcoin, in reality the capital costs of designing and manufacturing rigs will be quite high. Allot of resources will go into manufacturing, with the associated environment impact. Electronics manufacture requires rare, pure, toxic substances, precision industrial machinery, world wide shipping etc. If Moore's law holds for mining bitcoins, the old hardware might be tossed before it has a chance to equal the environmental footprint of it's production. This is certainly the case with consumer electronics.

You'd create a spread out 1000-core ASIC hashing wafer. 1m^2 x 1m^2 might do it. You're hashing rate is your temperature control and if you cranked it right up then the fusebox would melt. Your only limit is the amount of money you want to spend on electricity and the fuse current limit from your electricity supplier.

So let's assume you can make an electric hashing heater that's not too noisy and doesn't produce light - although that might be nice too - say 95% conversion to useful output. The question then is the relative price of energy sources.

Natural Gas - 4.1p / kWh
Electricity - 13p / kWh

* source http://www.biomassenergycentre.org.uk/portal/page?_pageid=75,59188&_dad=portal
** These are bulk fuel purchase prices.

Looking at http://www.decc.gov.uk
It would appear that domestic costs for England 2010 for typical amount of use is...
Gas - 3.5p / kWh
Electric - 10.4p / kWh

Assuming 98% efficiency of a condensing gas boiler or 70 -80% for conventional boiler. You could expect an all electric heater to cost about 3 times more than the equivalent gas boiler heater. Once you factor in the additional costs of manufacture etc then you're looking at getting less than 10% of the bill covered. When you also take into the account that these people might have environmental concerns about the tripling of their fuel bill (even though they'd only be paying a third and energy consumption was ~ the same but when's logic ever mattered to an environ-mentalist, right?) then it's starting to look like a non-starter.

I've convinced myself that this isn't workable but, the point is that their is a competitive advantage to measuring 'real' energy use by hash computation.

Bitcoin air conditioner anybody? I guess that's not possible either or you'd have servers farms that self-cool.

I'm still of the opinion that Bitcoin needs to measure real world 'proof-of-work' or it just won't succeed due to the huge demand in energy to run the thing.

If only there was some way to convert the hashing heat back into something more useful...

Quick! Someone invent a turbine that hashes as it spins!

Hmm... interesting stuff this Bitcoin. Makes you rethink all sorts of things.

[wareen]

To conclude from my perspective:
Yes, if price shoots up 200 fold within a year, Bitcoin's total energy consumption will be much higher by the end of that year but your conclusion of a 1:1 relationship is ignoring so many factors that it most certainly does not even hit the right order of magnitude.

This isn't mindless extrapolation. This is concentrating on the primary market driver.

I'm sorry, I did not intend to accuse you of anything mindless - your concerns are very valid!
I just think that the development of the Bitcoin economy is so hard to predict - even for weeks - that I think its just not feasible to make useful forecasts over years or decades.

In summary.. if bitcoin becomes a major world currency its price will soar -> mining rewards will soar -> amounts spent on mining will soar -> amount of energy consumed will soar -> number of power stations will soar.

Yes that connection is indisputable - as for the exact numbers and the timeframes our opinions diverge, so lets settle with that

Of course, if Bitcoin really does become a major world currency, the implications for large parts of the world economy would be very severe, making predictions of the overall ecological impact even harder.

One more point I'd like to add: the potential of Bitcoin to "sell electricity" without having a good connection to the power grid.
There are lots of remote places where electricity could be generated efficiently but where it's not economical because it is too far off the grid. A single wireless / satellite uplink is much easier to install than building a long-distance power line.

Regarding future mining hardware: I could imagine solar panels in connection with decent GPUs to be a long term competitor to FPGAs or even ASICs. They share high one-time costs and little or even no operating costs but with the solar+GPU combo you always have the possibility to earn money on selling electricity whenever Bitcoin mining stops being profitable (or even in the case of a Bitcoin collapse). If Bitcoin will ever play a major role in the world economy, Bitcoin mining will probably be tightly connected to power plant operators for its unique usage of electricity.

[ElectricMonk]

PS. I worked for many years on the thermal design of electronics in hot sandy places. I even made an evaporative cooling jacket for people who had access to water rather than batteries. I'd like to think that I'm not making this shit up

EDIT - removed quote from previous poster as it wasn't relevant to this comment.

[flug]

In summary.. if bitcoin becomes a major world currency its price will soar -> mining rewards will soar -> amounts spent on mining will soar -> amount of energy consumed will soar -> number of power stations will soar.

Yes that connection is indisputable - as for the exact numbers and the timeframes our opinions diverge, so lets settle with that

Yeap, I'll go with that. It was bugging me that I'd seen a potentially serious problem, and it wasn't being sufficiently acknowledged. As long as we're all aware of it, and factor it in to our predictions for the future of Bitcoin (i.e. Bitcoin is not infinitely scalable without cost), all is good.

[ElectricMonk]

If Bitcoin will ever play a major role in the world economy, Bitcoin mining will probably be tightly connected to power plant operators for its unique usage of electricity.

^^ True that.

As another commentator mentioned - technological improvements does not help the issue with the 1:1 price/energy in a wasteful mining system. If Bitcoin is to remain a proof of work then it needs to do something along the following lines...

Fuel -> electrical energy -> hashing -> electricity energy -> consumer.

The market will/would be driven by how efficient you can make the conversion to and from the hashing. The BTC price will reflect how much proof-of-work you can measure. The hashing seems a lot like a 0.01 Ω sense resistor in a circuit. You want it measure work without throwing too much away. Best place to put that sense resistor is probably somewhere within the power station as that's the only place you're likely to get any kind of efficiency.