OK, silly limitations aside, we need to start with some baseline principles:
1) ASICMINER should never have more than half of all hashing power. In order to make sure this never happens even by accident, we need to consider what happens if a large competitor suddenly has an outage, like what might be caused by a natural disaster. For example, if two players each have 40% of all hashing power, and either of them has an outage, then the other winds up with > 50%. In other words, the absolute maximum hashing power of ASICMINER can never be greater than all other competitors combined, minus the largest competitor.
ie: if Avalon is the largest competitor and they have 20% of the network, ASICMINER should never exceed 40% to avoid reaching 51% in the case of an Avalon outage.
Related to this, if we are considering the natural disaster scenario:
1a) We damn well better make sure that these large players are distributed throughout the globe -- a reason to seriously consider other ventures like AMC.
1b) We have to be able to use commodity equipment. It's critical to the maintenance of a decentralized network. Since we've switched to ASICS, this means ASICS must become so available that anyone can buy one (like a GPU).
2) There is a practical limit to how much people will pay for hashing equipment, and this is also bounded. Expectations for ROI on mining equipment are bounded by network growth rate. When network hash power is not on the rise, ROI can be long term. When network hash power is on the rise, 100% ROI must be obtained within relatively short order (before the equipment becomes obsolete). Right now the time tolerance for 100% ROI on ASIC hardware is probably around 3 months from when it starts mining. If network hash rate growth slows, we'll tolerate longer term ROI just as we did with GPUs -- maybe a max of about 1 year. If we zoom out to the big picture, we can probably derive a ratio of X BTC invested in mining must produce 100% return in X months. Right now mining returns are measured in BTC generated and mining fees. BTC reward keeps halving. Due to deflation (in terms of buying power), it's not likely that mining fees will rise per transaction, but we may experience some rise due to increased number of transactions. Still, there are only so many people on the planet, and ultimately only so many transactions that can go on (obviously we need to increase funding to our space program).
3) Due to the increasing buying power of BTC, we will likely experience more technological leaps in the coming years. But they will still be constrained by the same limits. Cost of electricity is indeed a factor, but I would argue that it is negligible compared to how quickly your equipment becomes obsolete. Remember that we're calculating in BTC, not in USD.
So, the short version:
1) we can estimate how much BTC ASICMINER can earn as a percentage of the network and we know that this figure is diminishing over time.
2) we can estimate how much BTC individuals can invest in mining equipment by using the relationship between network growth rate and ROI, and we know that this figure is diminishing over time.
3) the increasing buying power of BTC means that operating expenses aren't a major consideration
If these ideas are sound, it means we can come up with a maximum stock price for ASICMINER (or any other such company). It also means that we know that this maximum will decrease over time. The long term challenge is to get as close to that maximum as possible. The short term challenge is to convince organofcorti to do the calculations
edit: I am currently all in
, and the poolswitching can be done when the whole batch is up or sold and we're waiting for the new 200 TH/s batch to arrive (aka: end of june or july)) 
