Don’t buy a Terraminer
Why Bitcoin Mining is a Sucker’s Bet (right now).
Given the current difficulty, and the rate of increase of difficulty in finding BTC blocks, it now makes more sense to simply speculate on the currency than it does to mine it with consumer “plug-and-play” hardware.
Recently, Cointerra announced (and sold out) the first batch of their TerraMiner IV, which computes BTC hashes at a rate of 2TH/s. This is by far the most efficient way to mine BTC, at a current cost of $3/GH/s.
You can get some really good background on why this is a deal in this very good, and very in-depth article on bitcoin miners.
Now, when you plug all of the relevant numbers in (there’s a handy calculator that mostly works at http://minr.info/) it will tell you that this miner should break even in 7 days. The guys that make at Cointerra are very good technicians, and terrible economists or they are simply looking to profit off of the current BTC hype.
When evaluating any mining investment, the investor must evaluate whether or not he believe that the investment will be able to extract wealth equal to or greater than the cost of the initial investment. If we were evaluating gold, we’d need to look at the cost of the land, the equipment, and the labor, and see if we think that we’ll be able to extract an amount of gold greater than the value of the dollars we invest.
Now, with BTC mining, this calculation is actually much simpler. We have to look at the cost of the equipment, and the cost of the power to run it. Even better, since there is no labor (functionally) we can pay for everything we need in BTC. We would still need to figure the cost of power, but, as I will demonstrate in just a moment, that’s not going to come into the picture with the Terraminer.
At today’s prices (USD$620 / BTC1) the Terraminer represents a sunk cost of BTC 9.767 with a 90 day lead time (February Batch). In order to break even, assuming free energy, the Terraminer needs to be able to extract BTC 9.768 over its useful lifetime. This is unlikely.
If we look at today’s difficulty of 609M, that suggest that the Terraminer would be able to extract BTC 1.644/day. So, by that math, if we could plug it in today, we’d be happy in a week.
However, when we look at the difficulty number, it seems to have been doubling in somewhere between 15 and 30 days. Let’s call that a continously compounded growth rate of 3%, just to make the math easier.
If we look at today’s 609M difficulty, with a 90 day lead-time, you’re looking at a difficulty of 9B by the time you can plug in your miner. That leaves you generating BTC 0.1105/day.
If we further assume that we’ll be able to generate BTC at the highest rate for 30-day chunks of time, then we get a payback picture that looks something like this:
In this scenario, we’re looking at the device extracting a total BTC 5.58436 over its lifetime (as it will extract a number of BTC/day approaching zero), with a useful lifetime of about a year. After its first year of operation, assuming our 3%/day difficulty increase, it will generate roughly BTC 0.000000916170/day, which, to save you the math, is roughly $1/day, if each BTC is worth $1m.
If we adjust the growth of the difficulty rate downard to 2.5%/day, then we get a useful lifetime of about 15m out of the device, and an overall profit (!) of rought BTC 0.18 over the lifetime of the device, representing a 2% APY. It looks like this:
Further adjusting the difficulty growth rate down to 2%/ day, we’re looking at an APY of 38% over a 2-year device lifetime with a total profit of BTC 8.39550. Totally respectable, and also where I think that the current hype around any such machine comes from.
Unfortunately, there’s no reason whatsoever to believe that the growth in difficulty will plateau or slow. And when we take into account halving, I suspect this picture becomes much worse (but I don’t fully understand enough about BTC to know in what way).
I think that in the final analysis, if you have the capability to construct mining hardware, you have two options: a) sell it, or b) use it to extract wealth for yourself. Logically, any actor will do whatever is more profitable for himself (in the long or short term depending on outlook). The creators of mining hardware believe it to be more profitable to sell you the hardware than to plug it in. That should tell you something.
Now, there’s always something to be said for being ahead of the curve, and you will probably be able to resell your Terraminer for most of it’s value after you’ve extracted most of what you can with it (you’ll earn more in month one than any other month). However, at this point, you’re not mining, but performing currency arbitrage, where the currency you’re using is 2TH/s mining equipment, rather than BTC. This is similar to a financial derivative. If we examine mining solely on its ability to generate BTC, this is not a profitable strategy.
Given that better hardware drives a higher level of difficulty, which further drives a downward pressure on the ability of mining hardware to produce BTC, it should be possible to determine a manner in which the cost of mining hardware drives the cost of BTC in USD.
I don’t pretend to fully understand BTC, and I’d love to know where I’ve gotten this wrong. So, please, someone let me know! You can find the spreadsheet I used to create my findings here. Some have pointed to a simulation of the revenue to be generated by a Terraminer. The issue I have with this is mostly the fact that it figures the difficulty as progressing lineraly, where as recent charts seem to show a geometric growth.
Given that the currently most cost-effective miner will not be able to extract its purchase cost, the better bet at this point is to purely speculate on the BTC currency, rather than attempt to create wealth through mining.
