Jake Smith has another good article / overview of a large Bitcoin mining operation in China, “Inside one of the world’s largest bitcoin mines.”
The article states that this operation’s output is roughly 5% of the entire network hashrate and the electrical costs for it are about $1 million per month.
A couple guys on reddit did the math and came up with these other two numbers:
- $1 million electric bill per month = $33000 per day/$500(price of btc) so they need to mine 66 btc/perday just to cover electricity.
- If they have 5% of the network, then it’s 3600*0.05 = 180 bitcoin per day, or about $90k.
So they are generating $90,000 in revenue per day yet fully 1/3 of the costs are soaked up by electricity. Note: Last month, the bottom line price at a farm like this, to “create” a bitcoin was 2700 RMB ($444).
What this means is that if this is the most efficient set up possible (economies of scale via low labor costs, quick installation from manufacturer, relatively cheap land prices, relatively competitive electrical rates) then to power the rest of the Bitcoin network with similar data warehouses, the global cost for electricity alone would be around $240 million a year. Obviously this may not be the case as each geographic region and jurisdiction have several variables that could impact and move this final amount up or down, yet that is probably a relevant range.
Similarly, the hardware costs would likely double, triple or perhaps quadruple the costs as well. Add on costs of maintenance (things break), rent, etc. and this pushes the world wide costs of bitcoin mining upwards into the $1 – $2 billion per year range, which as copiously detailed in Chapter 3, is what theory predicts (MV=MC). This does not also take into account all the various exceptions to the rule of miners mining at losses to get “virgin” coins, or researchers externalizing costs onto government run computers, etc.
This dynamic could change as market prices for bitcoins fluctuate, see further discussion from Dave Hudson on this.