Bitcoin Forum

When one considers either buying ASIC hardware outright or investing in a mining company, it would be useful to have some idea of book value.  You can safetly assume within 18 months (likely generous assumption), almost any ASIC purchased now will be non-cost effective to continue running.  I believe if you purchase for example a BFL 60GHZ (I see one on ebay going for about 10k), the chances of making that back in 18 months just off of the coins produced is very little (even if we assume the exchange rate stays around 65BTC/USD which seems optimistic).  In order to make a reasonable decision that buying an ASIC now can ever lead to a worthwhile ROI (to me that would be about 25% in 18 months), the question is will these ASIC units have any reasonable re-sale value?

For example, my friend had a GPU miner with 5 7950s, which he recently shut down and has sold the 7950s (just over a year old) for an average of $230 each on ebay - considering they only cost $320 each to buy, that is a pretty minute depreciation of only 30%.  That is because these GPUs still have use in gaming and you combine a couple in Crossfire mode, and they can keep up with even the newest GPU cards and a lower cost (especially with some overclocking).

In order for ASICs to not drop to near neglible value, they would need to have some use after it is no longer profitable to mine bitcoins.  If the Fiat/USD continues to decrease and the hash rate increase exponentially, this time will likely come soon for the "older" (ie a couple months old) Avalon and BFL units currently in use - especially some of these units are not very $/Watt efficient compared to the newer ASICs being developed.  So, I guess the question is: are there any alternate valuable uses for ASICs once BTC no longer profitable.  If not, then you need to base your entire investment on the Bitcoins produced assuming hardware resale goes to near 0, which is a much bigger challenge to face than the GPU miners ever had to since they can sell their hardware for about 70% if the original cost.

0

Its not like they can draw pretty graphics is it  ;)

So people will be taking out the PSU and throwing away the rest?  It makes it hard to justify purchasing an ASIC if you cannot be guaranteed to get one of the earliest versions of the newest technology. 100% depreciation is pretty unheard of in most investments.

They have no other uses.  Everyone who is buying them factors this into their calculations.

I do not see how a BFL 60GHZ at 10k will ever make the 155BTC required to break-even - and that is assuming the usd/btc ratio does not continue to decline.  The only way I see to be profitable would be to pre-order the next "new" ASIC so you get it before others and then you of course take on the risk of whether the company will actually deliver on time or at all.  Not a great risk/reward ratio - seems buying ASICMINER shares a better risk then buying hardware directly.

They will still be worth a lot for a while, even when most people find the power costs to run them are no longer worth the return, because there are always people who have free power, or extremely low electricity rates.

We see the same things happening right now... people setting up litecoin farms in their free-electricity residences etc.

At some point, the owners (guessing mostly dorms) of these electricity "free" residences will notice the higher utilities and crack down on it. 

Dorms don't pay any attention to electricity.  Only way they would ever know is a staff member seeing a farm of gpu's.  And its highly unlikely most will have it against rules anytime soon.

Asic resale value is at a premium ATM.  Until they there are more on the market i predict they sell for higher then cost for a while.  Once more are on market it will drop dramatically, specifically some of the cheaper per GH ones.  Another advantage of GPU is they are not tied to coin value, if a big drop hits GPU's still most likely the same value on resell.  

Why do you think the economic lifespan of an ASIC is only 18 months.  Imagine for a second you bought a GPU 18 months ago and no FGPA or ASICs existed.  Would the economic value (not resale) of your GPU be $0.

You may say "wait people can make faster ASICS" and that is true but we are talking about a marginal increase not a magnitude increase.  Going to a smaller process means about 2x the electrical efficiency (MH/W) and maybe 1.5x the MH/$.  So people have more efficient rigs but no so much more efficient rigs that yours will not be able to cover the cost of electricity.

Space heater?  ;D

This is exactly what I've been saying all along, the prices they're charging for ASICs is too high to make payoff a reasonable prospect (except for ones delivered a few months ago, obviously). This is why I'm not going to be buying a bunch of them, if I can get a few for a reasonable price I might just for fun but the ASIC price vs expected revenue is pretty bleak. The difficulty curve is ridiculous right now, constantly increasing at a huge rate. GPU mining isn't profitable anymore and ASICs are very pricey right now with no resale value. The whole situation makes mining very unattractive if you run the numbers.

There are two things that can keep you mining at this point, goldrush mentality and being an early adopter. I have niether so I'm staying away. If the price of ASICs drops enough I may pick up a few for fun but that's it. Electricity costs + my time make it unreasonable for me to GPU mine right now.

I can think of one other use for the usb erupters, but i'd rather not mention it. i'll let someone else open pandora's box. :o

Could a researcher develop software that uses SHA256 hashing for something like Folding@Home to make use of cheaply available discarded ASIC units?

When I look at ROI, my 5 BFL singles arrived June 31st, 4 of them have already been earned back, so resale value isn't much of an issue.

Only if the distributed project relied solely on SHA-256 hashing operations. If it needed anything else the hardware would need to be redesigned or those operations performed on another device: CPU/GPU.

no, not really

ASICMiner USBs are going for $140-$200 on ebay and amazon.

Keep in mind that even after these units are no longer "profitable" they will still be generating coins at a lower cost than buying them outright for some time after that. Many people would consider this to be an added benefit, as they will generate "x" amount of coins profitably, and then "x" amount of coins at a cost lower than outright purchase for some amount of time. If you factor this in they have a longer lifespan and I believe they would likely save enough money to be worth it in the long run for someone that is just trying to generate coins over time.

I haven't actually done the calculations, but that is something I consider when I argue with myself over whether I should invest in expensive gear.

as long as the cost to mine, IE the electricity + time + environmental factors, is less than the value of coins generated there is inherent value in the ASICS.  Look at the difficulty required to make current ASICs "not profitable. You are crazy if you think people are going to keep adding to the network until the marginal gain is zero.

Regardless of rising difficulty, it's inevitable more competitors will enter the market. More saturation of ASICs is going to lower the resale value in a big way. I'd be surprised if an Avalon fetches any more than a few hundred bucks by year's end. Of course that's not a big deal to those that have them now, they have or will reach ROI soon.

Even if I just break even vs electric cost with my ASIC, thats fine as Im very long on BTC and fine holding them for years...

Mainly I mine to help secure the network which currently holds my coins; the fact that I generate a few coins here and there is just a bonus.

Scenario:

You buy 10 USB ASICs for ~11BTC.
You mine at ~3.3GH/s for 30 days and get 1 BTC
You sell 10 USB ASICs for ~10BTC.
You break even

Scenario 2:

You buy 10 USB ASICs for ~11BTC.
You sell 8 on eBay for ~$120/ea.
You use the $960 to buy 13BTC on Btc-e
You keep 2 USB ASICs plugged in producing ~0.01BTC/day for a few months
You win


Just depends on how fast your turn-around is and what you are purchasing and your intentions.  Strictly speaking, though, a GPU is going to have better resell in the long run.  After mining for a year on an ASIC, then something is going to be out that is going to dominate your ASIC. 

/end rambling rambles

You can get them for 1.05 BTC ($81.37), if you know where to look. I just want to accumulate some coin and hold for the long run. Its a hobby for me and might make a nice profit, no one can predict future BTC prices and diff is bound to level off at some pont. I have a couple of BFL units oddered April 2013 amd a full of USB Block Erupters on the way, mainly for shits and giggles.

I like the fact that the Usb Block Erupters are low power. I have a media server that run 24/7 anyway.
Its easy to set up a Linux or Windows VM to run them, or stop running XBMC on my RPI and use it.

I was thinking the same thing - but I made my move - maybe its just me because I made a lot from BTC. I finally sold all of gpus today, and there was a lot, 85000 khs in litecoin power. I got ~40k euro for all (I wanted to sell it fast). What I paid for all was nearly a 90k. But you cant look at it that way. I made a fortune (sold btc at 100-200 euro), so what I lost on price is miserable. The bigger problem is that I lost 2 years of my life, no social life at all, only bitcoin and fate in bitcoin.

So what I am doing now is that I got 2 avalons left, and I will be most likely investing all the money I got from cards to USB erupters (at first I thought 300 pieces, but seems it will be much more). The math is the following

I had 85 GHS so far, pulling 32kW of power. For same money I can get 800 pieces of erupters, making 260GHS, and pulling 2kW of power. So it is a big move. Even in a year, erupters will be worth something, not much, maybe 10 euro each, but till then, they will double the investment, maybe quadrupple it (money wise), and what is more important, I can have my life back, sleep, and I dont have to fight the heat....

One USB erupter costs 50 euro make a hashrate of 7850 card that cost 130 euro.  If you calculate a year for these 2, the erupter wins however you do it. The difference between 120W and 2W is too huge for the graphics card to overcome

As far as I can see by end of year any ASIC less than 100GH will be worth almost zero or maybe 1/10th to be nice as a resale value, diff will have sky rocketed beyond the moon

The difficulty won't rocket past the moon without the price heading back up.  Otherwise, who will be buying all the ASICs?  Why not just buy bitcoins?

The way I see it is everyone knows that when ASIC's are released in the wild in sufficient numbers come this Sep/Oct ( KNC, Terrahash, All Avalon chips orders, Bitfury and the 100TH setup, etc) there will a substantial increase in hash rate and the price of bitcoins will not increase as the ROI windows will be so small that most people and that includes myself will be dumping BTC daily to recoup ROI.

I hope I'm wrong as I've ordered ASIC and taking a gamble, if I break even at the end than I'm happy as its a hobby in sort...

I'm sure they'll be given new lives as spam/'ethical hacking'/folding boxes at some point.

Password cracking?

:-X

Not really good for that unless you know of passwords which are 640 bit strings and contain a 32 bit incrementing nonce. 

In theory you could make a "general purpose" SHA256 ASIC which would be useful for brute forcing passwords as well as mining but it wouldn't be optimized for either and this isn't the approach taken by any ASIC developer.

Ha! I'll just use it to compute my salt for pretty hashed passwords.

This is pretty generous, too. Most gains from dropping fabrication processes are not that big, though for some of the original partners who started at 110nm and 65nm, the gains moving to a much more modern process would be pretty big (and could be 2x) but once you start moving through the die shrinking process the gains are more in the fact they can produce more chips per sheet of silicon, reducing costs to make the chip in the first place. The gains in power and processing are not forgotten, but it's not always the main advantage of die shrinks.

I was thinking, Bitcoin ASICs are basically optimized cryptographic processors at resolving SHA-256 hashing problems.  Can't these ASICs be repurposed for the purpose of breaking strong encryption?

All it can do is calculate the SHA-256 hash (twice) of a 32 byte number.  So you'd need someone to encrypt their data with exactly that hashing function.

Application Specific. Afraid not :(. If they could, then their general purpose resale would be a lot more

THis is why I do not understand why ASICMiner sells for so much.  All their inventory once replaced by newer ASICs becomes nearly worthless.  And the dividend is fairly small.

There are many many potential ways to use proof of work to secure protocols: fight spam, dos attacks, etc. Any of these things could, if they wanted, use the Bitcoin POW function.

But few of them have been adopted— one reason proof of work (e.g. hashcash) hasn't been adopted for these things is because most of the time the enemy has a botnet and botnets are even better at PoW than desktop computers (e.g. the desktop user pays for power, the attacker does not).   But with a mining asic, this isn't so obviously so, at least for now.

So, if people want, they could go out an develop alternative uses for mining hardware. If some take of it may create a useful secondary market.

Their dividend is amazing!. Point me to one stock out there that consistently has >20% yearly returns from their dividends.

ASICMiner reinvests half of their profits in more mining hardware. They will never fall behind.

That requires two assumptions:
1. The fiat/Yuan ratio will remain stable
2. That noone else will come out with more efficient ASIC chips

Also, ASICMiner cannot be compared to a stock listed in fiat on a regulated exchange.  This is in bitcoin to an anonymous person "friedcat" eithe OTC or on an anonymous exchange.  The risk is much higher with no recourse if cheated. 

Nobody buys a share in a company for their inventory.  Nobody buys shares in Apple because they want 1% of Apple's office furniture.  ASICMiner has shown the ability to delivery AISCs in high volume.  Something nobody else has done to date.  Eventually more efficient processors will be produced however have you considered

a) more efficient processors don't instant-obsolete existing ones.  An ASIC on a smaller process in theory will have roughly double the efficiency (MH/$ and MH/W).  In reality it is much less maybe closer to 150%.  That doesn't make existing chips unprofitable .... just less profitable.

b) given ASICMiner's history it is plausible that when more efficient chips are produced ... they are produced by ASICMiner.

Yes, because efficiency is irrelevant to some people, because they have free or almost free power. Some countries have power that is almost so cheap its not worth metering, and in some places, they actually do that.

Also, where I live (Australia) there are a LOT of people who adopted solar when they were being offered lovely 200% feedback tarrifs. But with network saturation and policy changes these are lowering a lot in some places, and in the future, they may be 1:1 with the consumption tarrif and worryingly perhaps even lower. In this case, a clever miner could better utilise that power to run an ASIC farm. In other threads it has already been shown that with some moderate levels of ingenuity (usually involving a RPi and some erupters) you can even run mining systems directly from the panels themselves.

Of course, people living in dorm rooms and renters of properties that have included power / free power are more numerous and those people also won't care much for efficiency. They are essentially making free coins with that power, you could argue the ethics of this if you wanted to, but ethics usually matter little when there's money to be made and loopholes that can be jumped through.

I'd be willing to bet that even in 12-24 months time there will still be people around the place  mining with FPGA's and AM USB erupters, even if difficultly increases in the worse case scenarios. Yes, those devices might have to be sold by some people but they'll find new homes where electricity is cheap. Hell, I'll bet that some people will still mine bitcoin on GPUs: free heat and easier setup/tweaking than scrypt will attract at least a few people. And then, there are those who are into the conceptual, theoretical, and ideological aspects of bitcoin who don't care about efficiency or even profitability at all.