QFT



Dare I say, never? Well, never say never, but not for a very, very long time IMO.  I foresee a mid term future where the mining market is completely saturated and asic miners will struggle to operate profitably (and most will long have given up on earning back their investment). Under those circumstances, its almost impossible to warrant the multi million dollar investment needed for a new maskset unless you could achieve another "quantum leap" in power efficiency and production cost. Or bitcoin rises in value spectacularly.  

As will everyone else. But thats were the paradox here lies; if enough people "watch it closely" and dont invest, you might think that the ones that do buy them, should actually break even. But I dont think they will, because if too many people wait and see, BFL will just be enticed to lower their prices until they sell enough and the problem will remain exactly the same.

Its really quite an interesting economic case, but no matter how you slice it, in the end I cant see anyone other than a monopolistic vendor of the ASIC profiting significantly. Should a competitor appear around the same time, I can well imagine everyone ending up with a loss, including the asic vendors.

No, not at all. But profitability in the asic era will be determined mostly by the ones that are most retarded, and frankly, reading the forums they alone would pose a serious threat

On a more serious note, its not a matter of being retarded; you cant estimate the speed at which these things will sell and subsequently drop in price, and therefore the speed at which the difficulty will be going up. In a way, buying those asics at ANY price (well, above, say 5x marginal cost) is potentially "retarded". Its a bit like a dollar auction. let me quote wikipedia:

..a paradox brought about by traditional rational choice theory in which players with perfect information in the game are compelled to make an ultimately irrational decision based completely on a sequence of rational choices made throughout the game.
http://en.wikipedia.org/wiki/Dollar_auction

Buying ASICs is a bit like that. Except, you wont have perfect information, unless BFL opens its books and shows you the orders and backlogs and production schedule.
Did I mention a competitor yet?

Well, considering their minirig SC will sell for $30K, some of those assumptions make sense. But once difficulty has gone up enough to require prices per GH to go down by >10x or 50x, it will make equal sense to redesign for cost efficiency and not have 25 daughter boards with their own power circuitry etc.

Anyway, this discussion is missing the big picture. Even if you are right and prices would bottom out above $1000 per TH, thats still, what,  500x lower than today? And therefore, difficulty would be ~500x higher. Good luck estimating the ramp on that and earning back the investment of your minirig sc.

Current minirigs? Barely I guess. If nothing else, those heatpiped heatsinks arent exactly free.
The minirig SC? who knows. Lets wait and see if there really are 200+ asics in there, and how much or how little support logic etc they need; but if you doubt its at all possible to cram tons of asics on a PCB cheaply, look no further than some memory modules. They cost like a dollar more than the chips on them.

Regardless if 100x lower end user products are feasible, even if its "only" 20x or 50x it doesnt really change the issue. A miners ROI will still go up from, say, 1 year to his lifetime.

It doesnt matter one yota if transactions are processed by a a handful of  slow 100W CPUs or a gazillion 0.001mW ASICs. All that changes is difficulty, and arguably, security of the network, though that argument can go both ways.

To produce, chips are cheap as... well, chips. Assuming their 3.5 GH coffee warmer contains a single asic, and assuming its made on 130nm, and assuming my previous projections were roughly correct (1-10TH per wafer, which costs < $1000), variable silicon production cost of that chip would be in the $0.5 range.  Okay, thats a lot of assumptions, but you get the idea.

Obviously there is no point in trying to build a $1.5 dollar miner around that, put putting dozens of those chips on a PCB and putting dozens of PCBs in a case isnt rocket science and isnt going to cost 10x more than the asic itself.

I also wouldnt rule out the possibility these chips are not running anywhere near their full potential in the first products, but are instead capped, particularly in the coffee warmer.  Its what I would do if I were BFL.

Wrong. Prices of GPUs are relatively stable and predictable, since demand comes desktop users, gamers etc, same for FPGAs. Therefore bitcoin price/difficulty is currently fairly predictable.  Difficulty wont suddenly go up 10x unless BTC prices do.

ASIC is completely different. Its market value per GH is determined solely by bitcoin price/difficulty. And that difficulty will be determined almost entirely by ASIC price/$. See the feedback loop?  Then factor in the extremely low variable cost of an asic and perhaps you can see the problem.

There will not be an equilibrium until asic per GH market prices approach variable production cost, which is probably ~100x lower than these announced prices. How long that will take is what is completely unpredictable, unless you are BFL's CEO and no competitor will emerge. But if you think you can accurately predict price/difficulty 24 months from now, be my guest. I wouldnt dare bet if its 3x or 100x what it is now.

You keep failing to see the massive difference. GPU or FPGA pricing is not dependent on bitcoin price/difficulty.  If xilinx or AMD would somehow double or 10 fold their price/performance, you would expect a tenfold decrease in P/D. The other way around, there is zero impact. P/D skyrocketing for whatever reason will do next to nothing to AMD or Xilinx prices. IOW, GPU or FPGA performance/$ is a given and relatively constant, and it results in a relatively predictable P/D, not the other way around.

ASIC performance/$ is completely dependent on bitcoin P/D AND will completely determine it. This feedback loop combined with the enormous potential for price drops is what makes it completely and utterly different. 10x better MH/$ today will become 100x better tomorrow. And as a result, your ROI time will increase 10x too.


Im sure you can think of other ways to increase MH/$ than a $13 device.


Thats actually why I used a 24 month period, because Im also not expecting them to ship any asics this year. But 30%? Hey, cant let that pass by. Lets try this: 12 months after BFL shipped their first SC mini rig, the price/MH will be lowered by 50% or more per MH.

IM pretty sure both GigaVPS and Inaba would be smart enough to sell perpetual mining bonds on GBLSE. If they play their cards right, they will still make a nice profit, its the suckers buying those bonds that will be left holding the bag.

Gross margins are almost infinite; a while ago I estimated a single 130nm wafer would yield between 1 and 10 TH worth of ASIC chips. Such a wafer costs around $1000.

To answer your question; if they mine themselves, they only have to produce a few wafers to reach their maximum profitability, while not being shielded from the risks (BTC value, another ASIC surfacing etc).  It makes more sense sell many more ASICs at very high (gross) profit margins. Which btw, still wouldnt prevent them from mining themselves later, it would only take a few more wafer runs to out-hash anything they sold.

THis is somewhat akin to a dollar auction; miners can only lose.

Its rather the opposite. The fact they intend to launch at 10x better MH/$, while they could easily charge 5 or even 10x more considering the power efficiency gains, shows what kind of tectonic shift this really represents.  I expected something much close to FPGA prices initially. But these prices will only last until price/difficulty catches up, therefore asic sales would dry up, so BFL will slash them over and over until they are >100x lower than today and finally approach somewhere closer to marginal costs which is damn close to zero $ per GH.

That is the 'exponential' leap. Want to take bets that in 24 months, bitcoin difficulty wont be 100x higher than today?

Why is that silly? Its completely in line with academic papers and even a SHA256 research chip which promised ~40x efficiency improvement for SHA256 hashing from 65nm FPGA to 130nm ASIC.

Honestly, of all people, I am surprised you would not believe this.

AMD isnt going to be impacted. The entire bitcoin network consists of like ~20-30K AMD cards, most of them obsolete. AMD sells millions of cards each year, the overall AIB market is ~50 million cards per year. Granted, only a small % of those are highend cards, but this is still not something thats going to impact amd in any meaningful way.

I still want to short them though, but for completely different reasons.

But as I understand it, you wouldnt get the entry product, but the price equivalent of what you have now. IOW, if you would send in two FPGA singles you should get ~1 SC single @40GH. I dont see why you wouldnt take a (nearly) free ~25 fold increase in hashrate. It may be a band-aid on a wooden leg, but better than nothing and Id be surprised if it wouldnt at least cover shipping fees.

I said it would cost BFL far less than 1$ per GH (I actually calculated somewhere between $0.1 and $1 per GH on 130nm). Of course they are taking a fat margin on that, no one expected them to release these anywhere near marginal cost when they can charge 100x more. Over time however, their price will approach marginal cost.


Changing goalposts . I responded to this:

"40nm ASICs will do about 200 Mhash/Joule."

And I have no reason to believe this asic is 40nm. 65nm would surprise me, 90 or 130nm seems much more likely.

Not really. Ive said this so often and long before this announcement: people dont and wont fully realize the impact asics will have.

These prices look good NOW, until you do the math with difficulty going up 10x, 100x or more. For the record, difficulty went up 10000x over the past two years, and that was with off the shelve hardware who's price is mostly independent of bitcoin price/difficulty (AMD or xilinx dont adjust prices when difficulty goes up, nor can they charge 10000% gross margins just because miner would pay it). But a dedicated bitcoin asic will be priced almost purely by btc price/difficulty so expect its price to go down by orders of magnitude while price/difficulty goes up by orders of magnitude.

6 months after BFL started shipping their asics, the above prices will look like a terrible investment.


You'd be nuts not to. Have you calculated what your singles or mini rigs will earn you once difficulty reaches 10 or 100 million ?


No arguing that

Feature request: when you send an email for received (or sent) transactions, can you include the new balance of the wallet?

I would also prefer non html email.

Corrected .

Probably a lot smarter than buying their products. Still, remains to be seen if they are actually the first to hit the market. If they are not, they may have trouble recovering their investment.