Important note: this is not current information. Bitcoin and Litecoin mining have both changed a lot in the last few months, and this post is from April.
I spent some time looking into mining about three weeks ago - with the proviso that my hardware would be free, since I run a 3D animation company that's in the process of transitioning to GPU-based rendering - and came to the conclusion that right now, it still ain't worth it.
Was your conclusion that the power costs would outweigh the value of the mined bitcoins, or that the value would be marginal enough that it's not worth the bother?
A combination of the two. Power costs plus wear-and-tear on mission-critical components weren't outweighed by the tiny amounts of money we'd make.
To be fair, there's a bit of a conflict between the best rendering hardware - Nvidia, because of CUDA - and the best mining hardware - ATI. Obviously, I'd have been mining on NVidia. However, the numbers didn't even make much sense using ATI.
Bitcoin mining has been screwed because of the dedicated hardware, which has meant a flood of miners into Litecoin - which has proceeded to screw the Litecoin ecosystem too.
I shall keep an eye on the situation - I suspect there will be pockets of opportunity in the future. But right now, not so much.
Curiously, CERN made the same decision and came out the other way. For quite a time they mined Bitcoin using the spare CPU cycles of their data processing farms.
Mining was profitable to average users (i.e. not sitting on huge piles of cash) in 2011 and early 2012 probably (source: I owned two 5970s and some 5870s at mined quite a few coins with them). Right now you have to invest much money in ASICs (GPU mining isn't worth anything at this moment) which may also be never delivered or delivered so late that you're out of game by then.
Most people agree that if you believe in Bitcoin, the best way to invest is just to buy some coins and leave them in a safe place for a year or more. The economy is deflationary so if whole Bitcoin doesn't fail in near future, the price will eventually rise above the price at the moment you invested.
Is it deflationary because of the 21 million limit? That limit won't hit until 2140.
Or is it deflationary because the projected velocity will go up faster than the mining output?
It's deflationary in the sense that with passing time each coin is worth more than before. There is some correlation with money supply, but deflation ultimately means: how much can you buy for a set amount of currency.
The inflation/deflation of the US dollar for instance is measured by statistics like the CPI, which boils down to how much consumer goods you can buy. If you can buy 10 peaces of bread this year for $10 but only 9 peaces of bred next year, you have inflation. If you can buy 11, you have deflation. How the dollar supply (number of existing dollars) changes has a correlation to this (i.e. it influences this), but it's not defining it.
OK, so with passing time, coins are worth more than before because the usage demand is increasing faster than the mining supply.
I'm not sure if that is your argument though?
Essentially, yes. More precisely demand is increasing faster than readily available supply (people selling Bitcoins for goods or other currencies, also called liquidity).
When astronomical quantities of USD are being printed into existence (today's QE is like 80 bln dollars per month), CPI is not a good indicator. Rich guys who get all this new money cannot buy 100x more food and toilet paper. What they can do is to buy expensive property all over the world (not very liquid, thus not affecting everyday prices much) and stocks.
Right now CPI does not change as much as USD supply changes, so people don't worry. However, it does not mean this is not delayed and will hit down the road even if they suddenly stop printing new dollars.
GPUs? It's ASICs or not at all these days. Start with an FPGA prototype that can be used to create an ASIC, like an Altera. If you want to muck around with this stuff for fun you can buy a Cyclone or Stratix for relatively low cost.
It's an older article, but it serves to show just how fast these things swing. Just as CPU mining became worthless when GPU mining took off, GPU mining is totally useless now. Even with free power it's barely worth the trouble.
FPGA is still viable, the chips are fairly cheap and performance can be improved with better designs, but ASIC will absolutely crush it.
The good news is that getting your FPGA design taped out as an ASIC is actually not all that expensive, relatively speaking. Some will give you a few hundred samples for around $15K. Not cheap per-chip, but avoids the $250-500K setup fee you normally incur with most ASIC runs.
Yeah the main hurdle to getting an ASIC taped out these days is actually that most production lines are backed up. I think the sample route might be a good idea for what will hopefully be a small run.
They sell surplus area on their wafers to those making small runs, and the yields are not that great as these are the less desirable parts of the wafer. Still, it's actually quite cheap all things considered, especially if you only want a few hundred units.
The environmental ethics of this activity are horrible. Buy hardware that consumes lots of power and will soon be useless, and run it 24/7. I guess many of mans enterprises are just as unfriendly to the Earth but this one makes it particularly vivid.
I have a noob bitcoin question. Even after all the bitcoins are mined there will still be a need for "mining" for all the active transactions, correct? How does that work?
The block reward, which halves every once in a while, eventually hitting 0 after all are mined
And then the transaction fees. Basically if I want to send a BTC to somebody, I send the coin to them, and also sign over a small extra amount as a transaction fee. A miner who solves a block (ie, a bundle of transactions) can send those fees anywhere they like (their own account).
So in a long-run situation, mining effort will balance out transaction fees to come to some sort of equilibrium.
The average bitminer, weighted by the people who invest more, who'll be more likely to be better at cutting costs. The interesting part is due to both the increase in miners, and the reduction of bitcoin creation rate over time, existing bitcoins will always rise in price. Always be deflating.
> The increase in difficulty isn’t necessarily a bad thing, as the more difficult it gets to mine the blocks, the more people are willing to pay cash for bitcoins
That's the problem with his logic. Bitcoin's exchange rate has remained pretty much constant while the difficulty rate has increased 10x in the last 6 months.
Since the future price of Bitcoin is unknown, at the end of the day you can not really decide if mining is profitable with maths.
I stopped mining when I generated less money than I spent for electricity (the gear had already been paid off). However, back then the exchange rate was maybe 5$. If I had known that it would be 100$ two years later, perhaps I could have mined longer.
Even if you'd known that the exchange rate would eventually rise to $100, once the cost of electricity exceeded the current market price of the bitcoins produced the rational thing to do would have been to turn off the rig and use the money to buy bitcoins on the open market instead of mining them.
True, I just thought of that and came back here to add that :-)
What the Bitcoin experiment taught me is that I am not a very good investor... The good part of mining was that I did it for fun, so there wasn't much to lose. The mining rig is now my new gaming computer. So even if I hadn't made any money at all, at least I would have gained a new gaming rig. In that sense it was a good speculation, for me personally (no real downsides but potential upsides).
I felt that by mining I also supported the idea of BitCoin, so it had ideological value for me, too.
And possibly not with ASIC given the potential of multiple vendors worth of ASIC's coming online at roughly the same time. I believe at least one has shipped the bare hardware.
I've a little ASIC device sitting in the corner, but it's more a mild curiosity than any actual income. I've made a lot more money reselling the devices than I have mining with then.
Bitcoin mining is not a "job" (requiring effort) but an investment - requiring capital. So in the long run, it shouldn't be possible to get a return on bitcoin mining much above normal market interest rates, since if the return would be higher (as it was just recently) then an investor can buy/build large amounts of ASICs and drive bitcoin price down.
There's certainly people who have lucked out and made a considerable amount of money that would, for all intents, be a replacement job for them. Some of the early Avalon machines would have been making hundreds of dollars a day when they were received. Even earlier miners on CPUs would be rolling in hundreds of thousands of dollars.
They may have made a great investment with a large, quick return - but it's not a job because:
(a) the income from the Avalon machine likely is much less now than initially, and is likely to shrink rapidly - so it's not a sustainable future income;
(b) keeping the income at the same level can't be done by spending your time, but by investing lots more money again and again;
(c) you (and anybody else) can increase that income hundredfold by buying hundred times more machines - i.e., it is limited by capital, not by time/effort.
It sure is income, but with very different characteristics. It's even very different from classic 'passive income' investments like dividend stocks or rental real estate - such 'passive income' investments are with rather sure long-term returns; but bitcoin mining gear produces very volatile short-term returns, and is likely useless in the long run as the device you buy now will after a couple years likely earn less than the electricity cost.
That's in a literal sense. I meant mining as a replacement for having a job, not so much that the mining itself was a job. You're right in every way though.
I spent some time looking into mining about three weeks ago - with the proviso that my hardware would be free, since I run a 3D animation company that's in the process of transitioning to GPU-based rendering - and came to the conclusion that right now, it still ain't worth it.