I have see a lot of posts where folks argue that BTC price is somehow correlated with the cost to produce a bitcoin. That makes very little sense to me, at least as a matter of traditional economics. I have in the past invested a lot in mining stocks (metals), yet the prices of mining stocks bear little relation to cost to mine; rather, they correlate more directly with industrial requirements for metals (demand), major new discoveries of natural resources (supply) and to some extent the cross-correlation of each of those with economic conditions (recession versus boom times). I have a hard time seeing why bitcoin prices would behave any differently out of line with the proven laws of supply and demand.
Let S(t) be the supply of bitcoins (number of bitcoins in existence) at time t
Let P(t) be the price of a bitcoin at time t (for sake of simplicity, ignoring for now the bid-ask spread and issues of liquidity, which might come into play)
Barring any fundamental change in the demand curve (more below), it should be obvious that at any S(n+1) > S(n), that generally speaking P(n+1) < P(n). Why? Because if that were not the case, the future value of a bitcoin would always exceed the present value; which is illogical unless you believe that the present risk associated with a bitcoin is GREATER than the cumulative risk of holding the same bitcoin (assuming no changes in the demand curve, which is obviously the issue currently .. see below) and there is always risk of holding any currency (inflation, time value of money, and in the case of bitcoin regulatory risk). You can (by induction) prove this for all future times t + z.
So, if S(t) is always increasing (as it will for the foreseeable future), then P(t) should be decreasing at least a little bit.. unless the demand for bitcoin is increasing. The only possible answer is that the demand for BTC increases at a rate that is (far) greater than the dilution caused by the increase in supply. This is where things get really hairy though. Demand for BTC depends on what people are willing to pay for a BTC at the most fundamental level. This, in turn, depends on: (a) liquidity of BTC (ability to easily buy and sell with minimal transaction costs, low bid-ask spread), (b) ability to use BTC (retailers and merchants that accept BTC), (c) belief in the long term stability of BTC as a currency, (e) required use of BTC for certain people/transactions, and (f) speculation. The availability of BTC exchanges, coinbase, etc. has greatly increased liquidity. Publicity has increased speculation. A number of merchants (Overstock being the largest retail outfit) now accept BTC, which increases both demand and liquidity. A certain part of the underground (criminals, tax evaders, etc.) require BTC to remain anonymous (launder money and/or transfer funds out of the country without detection) which also creates increased demand.
Let D(P) be the aggregate demand for BTC at a price P.
If D(P) is constant between t and t+1, then P must go down over time (because S(t+1) > S(t)).
Certain aspects of D(P) are insensitive to P - e.g., criminals who have very few alternatives and would be willing to pay more for BTC than the rest of us. However, for the rest of us, D(P) is very sensitive to P and probably looks like a traditional supply and demand curve (greater price decreases demand, and lesser price increases demand).
So how does all of this relate to hardware costs?? The only relations that I see are inverse (e.g., would cause BTC to go down, not up):
(1) Increased costs (cost to produce 1 BTC, which is correlated directly with difficulty) **may** cause the rate of increase in rate of production of BTC to go down. Note, however, that the rate of production itself will still increase (number of BTC's produced in a given time frame will continue to rise); but increased cost/difficulty must eventually put a brake on the acceleration in the rate of increase in BTC supply such that the rate becomes linear rather than geometric. Still, however, S(n+1) > S(t) so the cost is actually inversely related to BTC price (more bitcoins should always produce downward pressure on prices; decreasing the acceleration in the rate of increase will put a limit on this downward pressure, but cannot eliminate it).
(2) Increases in P should cause more people to invest in mining, which in turn increases the rate of production of bitcoin. This should exert a downward force on P (S increases).
Perhaps the reason that BTC prices have been correlated in the past with hardware costs is that many (most?) manufacturers insist on or prefer payment in BTC, thus increasing demand prior to increasing supply (especially in the case of pre-orders). This effect, though, should disappear as the exchanges and greater liquidity enter into the picture (as they already have, I think) and the price of bitcoins increases to the point that manufacturers cannot unilaterally impact P by insisting on bitcoin payment.
The scary part of this is that when demand finally plateaus (as it must; otherwise there will be arbitrage with other currencies that will force demand to plateau), and D(t, P) becomes more fixed, then P(t+1) < P(t) as new bitcoins are minted all the time. In other words, when demand becomes fixed the price of bitcoins will start to fall as more bitcoins continue to be produced that increase S (until the maximum number of bitcoins are minted many years from now). This seems inevitable, but of course it depends on how long it takes for demand to plateau. In any case, I don't see how the cost of hardware plays any major role in all this.
