Maybe Asicminer will have gen3 units in-hand and ready to sell in the next couple of months?  Has anyone been following them?  Their hardware has always been expensive, but they (AFAIK) have always only sold units in hand, shipped quickly, and delivered exactly as promised meaning the ROI calculation is straightforward, reliable, and low-risk.  Of course, on the other hand, (at least based upon their share price), gen3 might be vaporware itself and not something to rely upon...

FWIW, with the slowdown in network size increases, I have been running my 65nm BFL hardware longer than I anticipated and I think COG's BFL hardware will be economical to run for a while longer (the returns are small, but they more than offset cost to run still I believe).

Seems like a motion is appropriate, but I can see this working a few different ways:

1. Sell hardware, use proceeds to pay 100% div.  Raise COG.F3 if we need more funds for hardware purchase later
2. Sell hardware, use 50% of proceeds to pay div, keep 50% for reinvestment fund
3. Sell hardware, put it all in reinvestment fund

I think 3 is only reasonable if we have some other hardware to buy that we can have in-hand very soon (like, < 1 month).  Otherwise, I'd prefer 1 or 2.  Would 1 mean "the end of cog", or would there still be plenty of money left over for reinvestment in the next big thing?  Maybe to answer these questions, we need to see just how many CT machines we sell, and how much profit we can make from them.  Also, if some of the CT machines do work, or if we are able to run a smaller number of them reliably, maybe we don't sell them all - just a few.

Also, maybe before we are able to sell them, the data center comes through and magically fixes everything.  Who knows.  I'm still in favor of keeping our options open, so long as movement happens to keep as many options open as possible in the meantime.