Quote from: Puppet on September 19, 2013, 07:16:53 AM

Do you think a for profit company that would start its design from scratch today, tape out next spring and have products next summer (at the earliest) stands a good chance of earning back its NRE? I very much doubt it. If a for profit company cant, I dont see why you think a cooperative can.

Why do I doubt it? We are now at what, 10 companies working on 28nm designs? You may be shocked by the premiums they are asking for preorders, but those prices are simply a result of myopic miners and the current low difficulty. Prices will just follow difficulty (+whatever reckless/clueless miners are willing to lose) and difficulty will follow asic pricing. Thats a feedback loop. The result is prices will plummet very very fast (as will profitability per GH). This would be true if there was just one supplier, but with a dozen that are all racing to lock in orders this will happen as fast as these companies combined can produce asics. Unless they are all even more inept than BFL, by the time you reach the market with a product, you will have to compete against 10 companies that have already recovered their NRE and are dumping chips and preorders on to a saturated market for prices barely above marginal cost, a tiny fraction of what they are today. You wont stand a chance recovering your investment.

Even if Im wrong about that, what could you possibly achieve? Assume you would be ready today, and you would be selling chips at cost (which means ~50x lower than current asic prices). What would happen? First off, you will be swamped by orders and you would never be able to fulfill them. Even if you could supply a massive amount, you would just push up difficulty tremendously, accelerating the scenario I described above. Sure, you would lower asic prices even faster than they would otherwise, but profitability would evaporate equally fast due to the D. You gain the community nothing.

There is only one "problem" right now with these asics; miners arent doing the math, or are deluded and are willing to pay far more than whats good for them. Thats the only reason prices are what they are. If miners wouldnt be willing to pay them, they would drop, simple as that. I dont see how you can fix this problem by producing yet another asic. Providing cheap asics wont solve it, it would just result in even more miners expecting windfall profits. In fact it might make the problem worse; if miners cant figure out today that $3/GH in January is a loss, how will they understand that $0.2/GH is a loss? And it will be a loss if you could supply those chips on time to everyone that wants them, since you havent solved the fundamental problem of miners overbuying, you just provide them another way to overbuy. And if you cant deliver on that demand, you will just become another BFL.

It is still profitable, since no one is willing to opensource the design of the chip. And I think it will be profitable for a long time...

chip_painter

Posts

Merits