Let's imagine that Bitcoin is a company whose sole value is in the outstanding value of its bitcoins -- I'll refer to it as Bitcoin's "market cap." Today, Bitcoin's market cap is roughly 6.5m coins * $18 == $117m. (ref: bitcoincharts.com)
Bitcoin relies on network power to keep its blockchain safe from what I will call a "takedown" -- here, what I mean is the addition of transactions that are in some relevant sense not real. It is widely understood that if there were a network player with roughly 50% of the network power, that player could potentially sabotage the blockchain in just this way.
The current network power dedicated to bitcoin mining is approximately 6.7 terahashes/s. (ref: bitcoincharts.com) It currently costs approximately $1000 to create a rig that can mine roughly 1000 megahashes/s -- let's say then that the cost of a megahash/s is at current difficulty $1.
By the application of math, we can see that it would cost roughly $6.7m to create a farm of computers that would equal the current power of the existing network. If you want to factor in some overhead for a data center and electricity, you can add a few percent to this cost.
If I am in the right ballpark with these calculations, Bitcoin is subject to a "hostile takeover" (to extend the corporate metaphor) or "takedown" scenario. This is because it would only take about 5% of the network's total value to sabotage it.
Granted that the normal forces of difficulty and increased network power serve to limit the feasibility of such a takeover over time. My point is simply that at this moment, because of the high value of the coins and the size of the existing network, it is feasible to destroy or sabotage it. My hypothesis is that equilibrium occurs for this system when the difference between the "takedown" value and the market cap are much closer in value than the current 20x spread.
I am interested to know whether anyone else has been thinking about this, and to what conclusions he or she has come. Thanks!