Topic: What's the lower bound on mining hardware over the next year? ($/GH) (Read 1293 times)

"How low can you go?"


It will go much much lower.

I guess we can scratch $5/GH/s off the list and $2.5/GH/s has been breached if you count hashfast's MPP.

The price of coin and the price per GH/s of the miners is tied to the network hash rate.  You can't say what the network hash rate will be without also making a prediction as to what the price of Bitcoin will be and what the price per GH/s of mining equipment will be.  I think you are spot on about the network hash rate in 2014 with the current price of equipment and price of Bitcoin, but it will be much higher if the price per GH/s drops a lot or if the price of Bitcoin goes up a lot.  They are all related although there may be some lag.

As far as falling over themselves to buy equipment, I was referring to what will happen if when the price of Bitcoin skyrockets.

I don't see anyone falling all over themselves anymore.  6 months ago, 3 months ago sure but a couple months is a long time in Bitcoin lands.

At the recent offerings
xcrypt - unlikely to have sucessful IPO
IceDrill - scaled back from 500 TH/s to 250 TH/s.
HashFast - still hasn't sold out of initial batch of 220 TH/s
KNC - never sold out of Oct batch.  Ended up cutting prices 30% for Nov delivery to stimulate sales.
Cointerra - announced 2 PH/s weeks ago and no indication they have sold out despite cutting prices 16% to $7 per GH/s.

Now this is all for delivery in 2013.   The math gets even worse in early 2014 when the network is >6 PH/s.  I don't really see >$10 GH/s flying off the shelf in early 2014.   Even if it does we are looking at 15 PH/s or so by mid year so late 2014 looks even worse.  Even if the exchange rate doubles, if hashing power quadruples then ROI% will look WORSE then then it does today and today sales are slowing despite falling prices.  As for exchange rate going to $1K per end of 2014 well that is just a pipe dream and anyone who thinks that should just buy BTC today as it is unlikely any miner at any price is going to return > 700% in a year.

If Bitcoin is $500 by mid 2014, and a person can earn 10 Bitcoin with the miner over a reasonable period of time, they will be willing to pay somewhere close to $5000 for it. If you can mine 10 BTC with a 500 GH/s miner and the price of Bitcoin is $500, then the price per GH/s a miner will be willing to pay will be ~$10. Miners haven't proven themselves very good at math so far. If manufactures want to shoot themselves in the foot by drastically reducing price per GH/s, they have the margins to do it but I think they will determine the price by what they think people are willing to pay, not what it costs them to produce the miners.

Competition could force them to reduce prices but buyers falling over themselves to buy at current pricing is going to limit how far they reduce prices.  If the manufacturers are as bad at math as miners are, you may be right and they will reduce prices right down to the point that they fail to make a profit and take the miners right down with them due to the higher network hash rate.

I personally hope it doesn't get to that.  They should reduce the prices slowly over time once they are able to deliver within a few days of a a buyer placing an order.  This will mean higher profits for them to develop the next generation and really it will mean higher profits for miners because the network will not grow as fast.

$500-$1000 Bitcoin and $1/GH/s just doesn't make sense to me.

Well even KNC batch2 (Nov) is down to $11.50 per GH/s.  By the time every company has paid off their NRE and delivered their first batch of hardware difficulty is going to be much much higher, and there likely is little appetite for Petahashes more of high cost hardware.  Unless the price comes down nobody is going to buy gear which is negative ROI% from day one.  Anyone offering $10 per GH/s in early 2014 is likely not going to see a sale, by mid or late 2014 it is going to look downright silly.

The marginal cost of production is closer to $1.  Given finite economical capacity and multiple competitors it is kinda hard to believe that 1000%+ margins will be sustained.

If you believe that Cointerra will be successful and if you believe that the price of BTC is going to stay under $150 then my $10 estimate wouldn't make any sense.

It depends on if you mean raw chips or entire system?

The raw silicon (excluding NRE) is on the order of $0.20 per GH/s.   I think people forget that high quality DC to DC PSU aren't cheap and neither are high current ATX power supplies.  Between the two you are looking at at least $0.30 to $0.50 per GH/s.  Then add all the boring stuff life labor, PCB, minor components, assembly, testing, cooling, cases, etc and I don't see cost being below $1 per GH/s and retail is going to be more than that.  Maybe $1.50 on the low end but $2 is probably more realistic.

There are already units offered at less than that.  Cointerra is $6.99 per GH/s.

$10 GH/s by the end of 2014.

I expect they will do as ASICMiner did with the USB toys - reduce price, when they have caught up on shipping and have more hardware to move.  As long as they're selling it for more than their costs, and they're selling it for slightly more than they believe it will make (which was certainly true of the USB Erupters), it's in their best interests to sell it for btc or fiat, whichever they prefer.

They'll keep on selling until it's not worth making them, then go about retiring.

The real question is... would be in the asic makers best interest to sell the "low" price, or just self-mine.

At current prices, soon (or already) mining hardware will become unprofitable and hardware companies will need to lower their costs prices. What do you think the lower bound will be in the next year or so?