- How will Gen 3 be divided between mining, franchising and hardware sales?
It depends on the users of our chips. It is a safe bet that we will be moving more to pure chip distribution. But if the edge of low cost on whole devices and deploying still exists in 2014 on our side, we will do more mining/franchising/hardware sales ourselves.
He said it himself. Notice even the word "pure".
I don't know why are you guys so keen on dismissing that.
I'm not saying it will be in this next deployment, but it will happen.
AM value proposition might be good as a vertically-integrated mining hardware manufacturer today, but when planning for the future and looking at the company numbers, things may be different.
So what are you implying, that FC is expecting and planning for a future in which AM can better design and more cheaply manufacture Bitcoin ASICs than a major player like Intel, AMD, NVIDIA or Marvell? Even the smaller players like Cypress are likely to outperform in terms of both chip efficiency and cost per wafer (and are likely to jump in sooner).
Bitcoin mining is within spitting distance of becoming a serious market for the big players.
I would suggest that it is actually in ASICMiner's interest that Bitcoin does not grow beyond it's current market cap for many years.
They have roadmaps that go as far as 25 years into the future and even more. They do not just watch BTC price grow, get curious, flip a switch and BOOM, they're shitting ASICs. It doesn't work that way. They are also a publicly traded company, not a one-man "I make all the decisions" show.
Besides, there is the fact that they are already making chips for >75% of the world's population. Why would they switch to what is, in comparison, a minuscule market such as Bitcoin?
Yes, it may not be minuscule in the future, but for Intel that would mean changing their whole game plan on a big if, and furthermore, they would have to change it years in advance to be successful.
AM knows this. By not taking the a quick bold risk (even if the CEO wanted; being a public company, they probably couldn't) and hopping onto the ASIC business Intel and AMD, have given Friedcat not only the early advantage, but also the assurance, that he has 5-10 years clearance to make sure he makes good use of that advantage.
If you're reasoning were true, Nvidia wouldn't exist today, or any other chip maker as far as it matters.
EDIT: Also, I remember reading somewhere sometime ago about Intel publicly dismissing getting into ASIC, but I cannot find evidence so don't quote me on that, may just be FUD.
EDIT2: After rereading, I realized this is in fact how all big tech companies fall. They get too big to the point where bureaucracy prevents them from adapting as fast as they need.
I work in this industry. The roadmaps you are talking about refer to their core businesses (logic for Intel), but that doesn't mean they don't do anything else. There are also industry-wide roadmaps for process nodes, technology inflections and wafer size transitions.
A company like Intel would do the market study, and if they liked how it looked they'd just set up a small experimental BU, which would crank out a design in a couple months and shit out a few million chips. If they failed to find sustaining sales volume they'd write it off and the loss would be minuscule, pretty much just a few months wages for a dozen engineers and a little bit of fab time at a prior-generation node.
Size also doesn't stop a company like Cypress, who will more aggressively go after small markets. Also, TJ Rogers probably loves Bitcoin, so that helps.
FC's team is competent, but I don't think they compare to the engineers at any of these companies - or the rapid access to cheap volume manufacturing at modern nodes. That's the reality.