Just for the record, the "old chips", which you can buy easily in bulk (assembled in to prefab PCIe cards by the thousands), only cost about $12,000 per 1.2Ths. (That is the OLD chips, which draw nearly 1200w).
It makes sense that the new chips, which are not that new actually, are selling for nearly half. (Which is obviously a "early buyer" promotion. I am sure these will quickly rise to $12,000 once they begin rolling-out to "small consumers", while remaining about $6,000 to bulk-buyers.)
But they do lack customer-service for "small purchases". They are obviously inundated with millions of time-killing questions from people who don't even have actual money to pay for the item. Just as BFL had to deal with. They are not selling customer service, they are selling machines. Either you have the money to buy it, and buy it, or you don't, and are just wasting their time, and ours. (That is only to the trollers.)
By the way, there are few people who are in the forums who purchased BFL's 1.5Ths miners, and few who purchase the other 1.2Ths modules which stack by the dozens into cages. Those are big businesses, the ones selling cloud-mining services and who own the majority of the 7000+Th on the network now. Be glad that they are willing to sell such small quantities to us, at these prices, or they will eventually take that away from us too. Not selling unless we buy 10+ 1.2 THs machines at once.
Nothing is stopping you from buying the old crap at 60-100Ghs for $6000, which is still abundant and people are eager to get rid of.
Link, for those who don't mind paying 2x for old technology now... In bulk... (Which is a 3-month preorder too. Just for the record.)
http://virtualminingcorp.com/shop1/index.php?id_product=23&controller=product#/fh_exp_case_16_pcie_slots-no_expansion_case/fh_256ghs_module_a-1_280th_1_200_watts/fh_256ghs_expansion_a-no_256gh_s_expansion_modulesSome good comments here ISAWHIM.
It's also worth noting that current chip tech fab is 22nm (introduced April 2012, actual mass production coming later) and Intel has market lead on those compared to any other chip fab, these shrinks take a few years for fabs to size / retool to production, 14nm will coming middle of 2014 but Intel will likely have 6 months to 1 year lead on everyone else. Searching the web you'll see 28nm for ASIC production is bleeding edge, so if they are delayed a little - don't be too surprised. Going from 65/55nm to 28nm is a large leap forwards for current ASIC production and it's not going to jump too rapidly further unless there's massive demand, market cap and demand increase of bitcoin producing h/w, so much so that reaches something ridiculous that this demand pushes chip fabs to spend millions/billions is retooling for better than 28nm...someone from chip fabrication could chime in here and clarify further or correct the information stated, which was gathered based on simple web searching and my tech background. What's likely true is that existing ASIC chip fabs are using some production % to produce these mining chips and unless demand is high expect delays as profit/production model establishes itself.
There's folks making more in trading recently due to rise and variability of trading against other currencies, so I don't think there's that much growing demand for mining h/w, plus the rapid difficulty rises and new startups trying establish a profitable business model is making it difficult to get hold of equipment. I hope things stabilize, but more so actual bitcoin trading volumes increase so that mining not just for bitcoins which is going down but rewards for trades actually means something useful versus the power being consumed. The rewards for mining from trades seems pretty much small to nothing, which is really disappointing, it would be good if these actually get published and tracked, anonymously ofcource like the number bitcoins being found.
We agree, ISAWHIM does have some good comments. And Naran is right when he states that going from a 65/55 to 28nm in 5 months is not easy for any industry. Our 1.2's and soon to be released larger miners will be dependent on 28nm chips and nothing more. You can build a 1.2th miner with bitfury, bu the cost to build is roughly - well too much and not practical. This market will really unravel over the next 6 months, and we'll understand the cost to produce/buy/ and develop chips as we move forward. A good factor on new players entering the production market depends on a various things, but one major one that everyone seems to overlook, is how ALT coins perform over the next 6 months. In general, this will determines the LTV of the customer, which will define major risk factors when choosing to enter the market or not. That's what the big guys are waiting for if you ask me.