At some point, the electrical cost of mining a bitcoin will be equal to the transaction value of the bitcoin, rendering to too expensive to accept the bitcoin in a transaction.
This is ignoring network costs as being part of the invisible glue that hold the Internet together.
So when is this date?
I can find a few computations on mining costs on line, but they are mostly concerned with amortization of equipment costs.
I do get the creepy feeling that the big money is actually in selling mining rigs.
The world economy is probably around 70 trillion dollers/year. Smallest possible transaction is (nominally) 1 cent, with 7000 tC (did I just coin teraCent?) in motion per year. Considering the amount of times a dollar can change hands in a day, transaction values are probably a order of a magnitude higher. there must be at least 5 Trillion transactions/year.
Each transaction requires how many cycles of computer time, and ergo, how many watt-hours?
Well, a 70 trillion dollar pool of transactions would pay for a lot of hashs at the rate of 1TH/joule...
This does not look supportable below 1 btc = $5. Now perhaps the bitcoin 8-fold subdivision trick might help!
Any way I look at this though, the btc adds a computational cost (e.g. a tax) to every transaction, which means that only Moore's law could save us in the long run.
What am I missing here?
EDIT:
Two years later, I ran across Christopher Malmo's article in Motherboard that approaches the idea from the cost of electricity standpoint, and comes up with some better figures than my back-of-the-envelope calculations.
No comments:
Post a Comment