Bitcoin: worth the electrons it's printed on?

I recall reading even before I had any interest in Bitcoin or Bitcoin mining that the cost of electricity was probably higher than the yield one could get by mining. So I never expected to set up a mining operation on my personal computer and get rich quick. I did what I recommend anyone do if they are interested in getting to know how Bitcoin works: I set up a mining operation to learn.

The online resources about Bitcoin are full of advice: don’t bother. There’s too much computing time necessary, electricity costs too much, you’re better off getting onto one of the exchanges and buying Bitcoin. This sort of discouragement is entirely appropriate for the investor trying to make a quick fortune with a nascent electronic currency, but not for someone who likes to get to know a technology before using it.

In fact, right now is a wonderful time to get started with Bitcoin because you can still set up a mining operation with your home computer and literally make money. Not a lot of it mind you, but with a decent graphics card you ought to be able to pull down a few cents’ worth each day. Not really enough to spend, but enough to look at and see that there is something there that could be spent. You don’t get into model rocketry because you think it will take you to the moon, but it’s a great feeling when you see that rocket head up into the sky, even if it only goes a few hundred feet.

Over the course of the last month or so, I’ve accumulated somewhere around ten US dollars worth of various cryptocurrencies.1 In addition to Bitcoin, there are a number of other lesser-known currencies competing for the space. Litecoin, the second largest cryptocurrency, is touted by some as the silver to Bitcoin’s gold. What it promises is faster confirmed transactions and an algorithm that circumvents some of the ways miners are taking over Bitcoin production and squeezing out smaller players.

Most cryptocurrencies are much much smaller and have been created for short-term profit rather than a desire to innovate. Speculators often jump in on new currencies and buy up a lot when it is cheap, then dump it as the price increases due to the speculative demand. It can be very profitable to create a new currency, mine it like crazy, and sell what you’ve got early, leaving a bunch of speculators holding the bag. However, a few of these currencies stand out as having real potential:

  • Bitcoin, of course. The original.
  • Litecoin, which trades at very favorable rates, currently a bit shy of three dollars per Litecoin.
  • Feathercoin has some popularity among miners but I’m skeptical about its future.
  • Peer-to-PeerCoin, which adds a concept called proof of stake in addition to the proof of work that most other currencies rely on exclusively.
  • Digitalcoin, one of the newer entries into the field. DIgitalcoin’s developers have gone the extra mile with Digitalcoin, setting up exchanges for purchasing real-world goods with Digitalcoin. Most alternative cryptocurrencies exist only as trading chits by which miners can trade in easier-to-mine coin for the more difficult to create but easier to spend Bitcoin. It remains to be seen what will happen with Digitalcoin, but I have to give credit to the developers for doing what it takes to make a currency stand out and succeed.

I have small holdings in each of these currencies, all together totaling about ten dollars as mentioned above. Ten dollars isn’t a lot, but hey, free money! Right?

Well, no. Now we get back to where we started. Electricity isn’t free and processing power uses electricity. So how much did my ten dollars cost?

There’s a great device called a Kill-a-Watt that plugs in to a wall socket and which any appliance can in turn be plugged into. It has a digital readout of how much electricity is being used at any given time by whatever is plugged in to it. Its sole purpose is to measure electricity usage so that people can find out where they are wasting electricity. I used this Kill-a-Watt to measure the power consumption of my computer at (relative) rest, and then set it to mining.

I measured the power consumption in a few different ways, but all variations on the following:

  • Mining with CPU, yields approximately 20¢ per day2.
  • Mining with NVidia GeForce 8800 (a pretty bad card to mine with), yields approximately 5¢ per day3.
  • Mining with ATI Radeon 5770 (not great for mining but about ten times faster than the NVidia), yields approximately 40¢ per day4.

Here are my power consumption numbers:

Work being done Power consumption (watts)
Booted up, normal operation, no mining 340
CPU mining only 420
GPU mining with NVidia GeForce 8800 only 395
GPU mining with ATI Radeon 5770 only 415
Radeon and GeForce simultaneous 470
Radeon and CPU mining simultaneous 510
GeForce and CPU mining simultaneous 450

From these numbers, we can see that CPU mining adds about 80-100 watts, Mining with the GeForce 8800 adds about 40-50 watts, and mining with the Radeon 5770 adds somewhere around 80-90 watts. Multiply each of those numbers (lets just take the low end for the sake of unrealistic optimism) by 24 and you get the number of kilowatt hours I’m charged each day for using each of the mining methods above:

Device kw/h per day cost per day at 19.423¢ per kw/h approximate daily yield difference
CPU 1.9 37¢ 20¢ 17¢ loss per day
GeForce 8800 0.85 19¢ 14¢ loss per day
Radeon 5770 1.9 37¢ 40¢ 3¢ per day gain

This suggests that with minimum reasonable hardware (A GPU capable of 150 or more million hashes per second) you might make a small profit if you’re lucky on your power consumption. And (this is a very important caveat) if you don’t count the cost of keeping your computer running all day. Most people’s computers won’t guzzle quite as much electricity as mine does, but it is still a very real consideration, and more than enough to offset a three cent gain per day.

There are GPUs out there with much faster hashrates and more favorable power consumption, but it seems very clear: the electricity is likely to cost more than the mining will yield unless very close attention is paid to the hardware choices. While I recommend that anyone who is curious get started mining, I don’t suggest that you keep doing it for very long unless you have very cheap electricity or are willing to take a loss for the fun of creating your own money.

  1. Called cryptocurrencies because the technology they are built with is cryptographic in nature. ↩︎

  2. Using the scrypt algorithm used by Litecoin, currently a more profitable choice for those with lower processing power. ↩︎

  3. Using the SHA-256 algorithm used by Bitcoin. The comparison with the above revenue from scrypt-based currencies is therefore pretty unfair. However, I was never able to get scrypt mining to work on a GPU so I have to work with the numbers I’ve got. ↩︎

  4. Echoing the last two notes, but adding: none of these numbers will be meaningful for very long. The prices of currencies and the difficulty of solving blocks varies on a daily basis, each of which changes the profitability. ↩︎