Will any U.S. Manufacturer deliver 20nm this year?

[bitndx]

I was wondering what the community's read on whether any U.S Manufacturers will deliver 20nm products this year or are they going to limp it out on their existing technology?

[Collider]

I was wondering what the community's read on whether any U.S Manufacturers will deliver 20nm products this year or are they going to limp it out on their existing technology?

I think it is unlikely that any US Bitcoin ASIC manufacturer is going to produce a 20nm product this year.

Furthermore, it is unneccesarry, as there is much more improvement to be achieved on 28nm anyway.
28nm chips can easily be used to build 0.4W/GH devices, as you will see in the very near future.

The important thing for Bitcoin miners is power efficiency. Bitfury gets 0.785W/GH on 55nm devices.

This shows that in the mining world process node isn´t everything.

[cryptomatt]

KNC will be shipping 20nm in the Neptune supposedly starting this month.  So to say that no manufacturer will isn't accurate but it isn't looking like a "US" manufacturer will.

[bitndx]

I was wondering what the community's read on whether any U.S Manufacturers will deliver 20nm products this year or are they going to limp it out on their existing technology?

I think it is unlikely that any Bitcoin ASIC manufacturer is going to produce a 20nm product this year.

Furthermore, it is unneccesarry, as there is much more improvement to be achieved on 28nm anyway.
28nm chips can easily be used to build 0.4W/GH devices, as you will see in the very near future.

The important thing for Bitcoin miners is power efficiency. Bitfury gets 0.785W/GH on 55nm devices.

This shows that in the mining world process node isn´t everything.

KNC claims 20nm with .7w/gh. Based on what you are saying though that will not drive focus on upgrades elsewhere. If we allow that no US manufacturer will go 20nm this year. Will they even innovate enough to improve power efficiency like you state or do they really only have enough resources to sell their chips with current performance(hashfast, cointerra etc.)?

[Biodom]

I was wondering what the community's read on whether any U.S Manufacturers will deliver 20nm products this year or are they going to limp it out on their existing technology?

2 out of 3 US manufacturers are unreliable (BFL, hashfast). Don't know about Cointerra-lots of complaints there as well, but at least they produced close to the schedule.

[bitndx]

I was wondering what the community's read on whether any U.S Manufacturers will deliver 20nm products this year or are they going to limp it out on their existing technology?

2 out of 3 US manufacturers are unreliable (BFL, hashfast). Don't know about Cointerra-lots of complaints there as well, but at least they produced close to the schedule.

Cointerra says in there recent blog post about processing refunds that they will continue to innovate and lead the industry in both Bitcoin mining hardware design and Bitcoin cloud-based mining contract solutions.

I wonder what they mean by innovate since they seem to be operating just above "life support" stage.  Hashfast is on life support, but still selling chips and has outsourced board production...I don't see opportunity to innovate there. BFL, AMT don't seem to be able to deliver at all.

innovation seems like wishful thinking right now when survival is unclear.   

[DrG]

BFL will be releasing their King Monarchs in 2 weeks and those will be 20nm - and they'll help you get a girlfriend and lose 30 pounds and make your teeth 3 shades whiter.

Seriously, looking back at the track record has there been a decent US based ASIC manufacturer?  Avalon and KNC and Bitmain have proved to be leaders at one point in time irrespective of how unscrupulous they may be.

[Beyondo]

I doubt any of them will release a 20nm this year.  Just getting efficient 28nm stuff has been hard for most of the vendors.

[Collider]

Sorry, i was half asleep when posting, obviously i meant US manufacturer, as stated in the topic.

My point still stands though, 20nm first of all is not optimal for Bitcoin asics, as it is only meant as an intermediary half-node on the way to 16nm.

Second, there is much optimisation to be done on 28nm and like US car companies (the big ones anyway), except maybe tesla (which is tiny), i doubt that any US manufacturer will come up with an efficient solution.
Maybe BFL really gets their claimed W/GH though, so i might be wrong on this one.

Spondoolies seems to be on a good path with their sp30 solution, which should get 0.45W/GH at the wall and is also professionally engineered.

KNC should get 0.6W/GH, but their price again is ridiculous, their business ethics are also highly questionable at this point.

What i am definately not wrong about is that both the (now bankrupt) HashFast and the always-late BFL shouldn´t be trusted in this game.
Their retail prices for home miners are furthermore outrageous, ROI with these is impossible.

Again,  their business ethics are nonexistant, so i wouldn´t order from them.

It seems like anywhere else in the world Bitcoin ASIC designs can be completed more or less on schedule, so i really hope these "bad" companies will go bankrupt in the longterm.

[ThomasCrowne]

My bet is we probably won't see any actual 20nm boxes hitting anytime this year.  But by the middle of next year it's almost a certainty.

[Collider]

My bet is we probably won't see any actual 20nm boxes hitting anytime this year.  But by the middle of next year it's almost a certainty.

As already pointed out, knc is currently delivering 20nm miners.

However, they are terribly overpriced and get "only" 0.6W/GH, a figure that can be achieved on 28 nm.

[DrG]

My bet is we probably won't see any actual 20nm boxes hitting anytime this year.  But by the middle of next year it's almost a certainty.

As already pointed out, knc is currently delivering 20nm miners.

However, they are terribly overpriced and get "only" 0.6W/GH, a figure that can be achieved on 28 nm.

I think he was answering that no US manufacterers will be delivering 20nm (KNC being Swedish).