This was a post I drafted for another thread but works here – It needs some more work as it is very rough, I suspect the proportions of mining power are more skewed towards to 55nm and 28nm chips now.
I suspect the energy calculations for Bitcoin are pretty off on Blockchain.info:
- Electricity consumption is estimated based on power consumption of 650 Watts per gigahash and electricity price of 15 cent per kilowatt hour. In reality some miners will be more or less efficient.
The exponential rise in hashing power from 400 TH/s to 4000 TH/s has been largely powered by ASICs. Albeit a mixture of 110nm to 28nm… For the sake of comparison I’ve assumed a fairly even split between 110nm (Avalon, ASICMiner), 50nm (BFL) and 28nm (KNC, Cointerra, etc) in the last four months and I’ve taken figures from my own KNC Saturn (270 GH/s), Avalon 4 Module (100 GH/s) and BFL SC (60GH/s).
Saturn: 1.3 W per GH
Avalon: 9.2 W per GH
BFL: 4.6 W per GH
New hashing power: 3600 TH/s (3,600,000 GH/s)
Saturn: 1,200,000 GH x 0.0013 KW = 1560 KW
Avalon: 1,200,000 GH x 0.0092 KW = 11040 KW
BFL: 1,200,000 GH x 0.0046 KW = 5520 KW
Total: 18120 KW
18120 KW x 732 Hours = 13263840 KWh per month x 12 = 159166080 KWh per year
Cost @ $0.15 per kwH = $23,874,912 per year
159166080 KWh per year x 0.44548 KG CO2 = 70905305.3184 KG of CO2 per year for the Bitcoin network
This is the same as 23,396 cars on the road (Using data from http://www.epa.gov/cleanenergy/energy-resources/refs.html#vehicles) or 261,162 barrels of oil being burnt
This isn’t going to be accurate, I’m going to try and refine it though, as it only takes into account the new hashing power in the last four months (90% of the current hashing power) and I suspect the proportions are off as I imagine the 28nm and 50nm miners have a higher proportion than 33% which would reduce this further. The original 400 TH/s has largely been replaced by ASIC power I would guess, last time I looked GPU and FGPA weren’t profitable to run and it would be easy to have had 10% of the network power replaced.
So the Bitcoin network is now 130x times more efficient than it was just four months ago if we extrapolate the same 3,600 TH/s out at 0.650 KW per GH.
The real question is how much has this improved the network, reduced confirmation times and how much of that original 110 and 50nm ASIC power will move over the Peercoin (or gets shutdown).
If the price of BTC/USD levels out then I think we will see the older generation miners become operational unprofitable around Jan/Feb, however if we continue to have 50-60% rises in the BTC/USD then they could stay around until May or June. At this point it is interesting:
If a miner is looking at it from a fiat income point of view they will continue to mine on the older 110nm ASICs if the price of USD is still rising even though it would be more profitable long term to spend the monthly running costs of the miner on buying Bitcoins
If they are looking at it from a pure BTC income point of view then you buy BTC with the running costs and shut the miner down. Or move it over to mining on Peercoin depending on block reward and exchange price at the time.
What I think will happen
The cost of mining hardware is still too high. It is priced in USD and with a Bitcoin conversion the ROI are so slim it is less risk to buy that equivalent BTC (some are around 2-5% ROI).
A lot of mining hardware will continue to be sold though over the next three months. The price rise of BTC will still get people spending USD to be part of Bitcoin by buying mining hardware as they heard that ASICs make lots of Bitcoins (especially with the new press coverage).
We may start to see a leveling of the difficulty and hashing power around May to July 2014. As the difficulty rises mean that everyone who bought a miner is either trying to sell it second hand
[Space to insert some figures about peercoin energy usage and PoW/PoS comparisons]