HashFast announces specs for new ASIC: 400GH/s

[jjiimm_64]

...
I had collected a number of questions the days before, mainly around how to deal with MPP. HF is acutely aware of the issues a delay is causing and they were certainly interested in exploring what can be done with MPP to help customers succeed as that is the original intent of the program in the first place. I suggested approaches such as starting the timer on Nov 1st and/or accelerating the release of new MPP modules to compensate for delays. etc. They have to operate under an number of constraints so there is no clear solution they could share with me at this point but I'm confident they will work something out.

...

The intent of a business is to make more money then expenses as much as possible.  If you think any business has some sort of different charity intent then I got a bridge to sell you.  It might even hash or some thing like that.  There is no solution to their desire to make more money.  They advertised MPP to collect more sales.  It had zero to do with wanting customers to succeed because any ASIC seller wants to sell their ASIC for the maximum amount possible.  Your visit reminds me of in-person accounts of visits to Enron.  Lots of fluff.

Any company that consistently screws their customers will not be a company very long (unless you are a monopoly or BFL apparently).

While I agree that HF is a company and companies will try to maximize their profits I think reality is more nuanced than that. There needs to be a balance. They need to provide value. They need to build a customer base. They need to build a reputation. But thanks for your constructive feedback.

there is a simpler explanation:

Keeping customers happy DOES maximize profits in the long run.

[1714114251]

1714114251

[1714114251]

1714114251

[el_rlee]

"Mining Bitcoin" and "long run"

[yuriygeorge]

If refunds are granted for pre-Aug 15 Batch 1 customers, does it include the option to refund the Upgrade modules as well for those that purchased them?

[crumbs]

...
there is a simpler explanation:

Keeping customers happy DOES maximize profits in the long run.

Keeping customers happy is not necessary as long as the customers keep buying.
As the entire history of bitcoin ASICs shows, they're never happy & they always keep buying.

[iCEBREAKER]

So, I had a chance to visit the HF office in SF today where I met Amy, John and a couple of other folks including an ASIC engineer.

HF recently moved from their San Jose office (with Uniquify) to SF. Currently they are working at a co-working space called NextSpace near Potrero Hill. But this is only temporary as they will be moving to a dedicated HF office near Market St. & Bush St. I asked about the size of the company and If I understand correctly there are now ~14 HF employees covering everything from ASIC design (Mainly done in Australia) to Marketing, Sales and Support.

After a quick tour of the area we moved to the conference room where I chatted with John, and were later joined by the ASIC engineer (not Simon), Amy and another friendly lady whose name I forgot (sorry). Cara had warned me ahead of time that I would have to sign an NDA before even getting into the office, but luckily that was not necessary.

I had collected a number of questions the days before, mainly around how to deal with MPP. HF is acutely aware of the issues a delay is causing and they were certainly interested in exploring what can be done with MPP to help customers succeed as that is the original intent of the program in the first place. I suggested approaches such as starting the timer on Nov 1st and/or accelerating the release of new MPP modules to compensate for delays. etc. They have to operate under an number of constraints so there is no clear solution they could share with me at this point but I'm confident they will work something out.

Another topic was the possibility of refunds and John was pretty specific that if refunds do happen they will be full refunds in the currency the order was paid in (BTC in BTC, USD in USD). Please note that I have not requested a refund and have no plans to do so at this point.

John was also asking if miners might be interested in cloud hosting provided by HF (at industrial energy rates). As a residential miner that will probably end up with at least 5 modules after MPP and pushing the limits of my fuses at home I would probably make use of this if the terms are good. I don't think they have a concrete plan yet but if there is enough interest this is something they will likely pursue in the future.

Another thing we discussed is hardware, however since they are in a temporary co-working space at the moment with most of the engineering & manufacturing done off site (Australia for ASIC design, TSMC for the semiconductors, Ciara for the cases, etc.) they did not have physical hardware on site. John mentioned they are not allowed to take pictures in the factory, but it should be possible to get some pictures of the PCB. So I expect they will have something to show soon. Amy also had some older video of running FPGA simulations (from June) that she could probably share here if there is interest. (it was just a 7s video of a running FPGA board, so not very representative of the current stage of development imho)

As to the duration of the delays, I was unable to get concrete shipping dates. However, Amy did mention they did make significant progress on the substrate issue in the last 48 hours. But I don't have specifics.

Overall this visit exceeded my expectations. I was planning to stop by for only 20 minutes or so and chat with Cara, but I ended up talking with the entire crew for about an hour. My impression is that HF has a capable & professional team, and they do listen to their customers. Amy is also a frequent visitor at the Sillicon Valley Bitcoin Meetup every Tuesday at the plug and play tech center in Sunnyvale so that is another good opportunity to chat with HF if you live in the area.

Note: As I mentioned earlier, I did not sign an NDA but I also want to avoid spreading misinformation due to possible misinterpretation on my side. So HF if you are reading this and find anything to be incorrect, please let me know I'd be happy to fix it.

Thanks for the report.

Good to hear the adults have arrived.

Please ignore all the 'nothing has changed because I'm still whining' idiots.   

[cedivad]

... Said hf's biggest puppet.
Sorry, but i couldn't resist.

[Micky25]

Good to hear the adults have arrived.

Please ignore all the 'nothing has changed because I'm still whining' idiots.   

Meanwhile, in a different reality

Gives me some confidence...though I really wish they would update us regarding this tapeout...

The tape-out was supposed to be finished the first week of October.

It's now the last day of October.

Cointerra is, at minimum, one month behind schedule.

And of course they will have other delays.

So there is zero chance of Cointerra shipping in Dec.

Jan 2014 isn't looking good either.

So much for Cointerra's 'Dream Team.   

[ImI]

Good to hear the adults have arrived.

Please ignore all the 'nothing has changed because I'm still whining' idiots.   

Meanwhile, in a different reality

Gives me some confidence...though I really wish they would update us regarding this tapeout...

The tape-out was supposed to be finished the first week of October.

It's now the last day of October.

Cointerra is, at minimum, one month behind schedule.

And of course they will have other delays.

So there is zero chance of Cointerra shipping in Dec.

Jan 2014 isn't looking good either.

So much for Cointerra's 'Dream Team.   

LOL! Thats exact the point of the bullshit thats going on in this hardware manufacturers threads.

[iCEBREAKER]

Meanwhile, in a different reality

Gives me some confidence...though I really wish they would update us regarding this tapeout...

The tape-out was supposed to be finished the first week of October.

It's now the last day of October.

Cointerra is, at minimum, one month behind schedule.

And of course they will have other delays.

So there is zero chance of Cointerra shipping in Dec.

Jan 2014 isn't looking good either.

So much for Cointerra's 'Dream Team.   

Same reality, different ASIC company.

Cointerra has been shown to be all bark and no bite. 

Their heavily hyped 'ASIC All-Stars' have been stuck behind schedule, fixing a flawed chip design, all month long.

HashFast completed mock and actual tape-out in record time (to TSMC no less) back in August.

Cointerra is still struggling to complete theirs, four weeks after the target ("first week of October").

There is a 99.9% probability of HashFast shipping before Dec 31.

There is a 0.01% chance of Cointerra shipping before Dec 31.

[ASIC-K]

both hash fast and coin terra will be at least a month late, so they both fail horribly. MPP, what a crock of shit.

[iCEBREAKER]

both hash fast and coin terra will be at least a month late, so they both fail horribly. MPP, what a crock of shit.

By this point, anyone paying attention to the ASIC space knows to expect delays and discount wildly optimistic marketing babble about earliest possible dates given only the constraints of physics.

The main difference between HashFast and Cointerra is that HashFast's chip taped out in record time to TSMC, the most advanced foundry in the entire world, while Cointerra has been stuck behind schedule all month fixing their botched tape-out to Global Foundries, a crappy second-rate fab that survives on TSMC's leftovers.

Another BIG difference is the global hashrate/difficulty will be  about 6-10 times higher when Cointerra ships in Feb/March than when HashFast ships in Nov/Dec.

Cointerra spend their time creating slick marketing materials, hyping their underqualified Dreem Teem, and gathering funds, while they should have been working on their chip.

HashFast designed their chip in stealth mode, only emerging and beginning marketing efforts once it was done.

The contrast between the two approaches (sell first design later vs. design first sell later) speaks volumes about the companies' respective characters.

[User705]

both hash fast and coin terra will be at least a month late, so they both fail horribly. MPP, what a crock of shit.

By this point, anyone paying attention to the ASIC space knows to expect delays and discount wildly optimistic marketing babble about earliest possible dates given only the constraints of physics.

...

Why?  If anything you'd think by this point companies would know that delivery date is one of the most important issues for miners.

[User705]

...
I had collected a number of questions the days before, mainly around how to deal with MPP. HF is acutely aware of the issues a delay is causing and they were certainly interested in exploring what can be done with MPP to help customers succeed as that is the original intent of the program in the first place. I suggested approaches such as starting the timer on Nov 1st and/or accelerating the release of new MPP modules to compensate for delays. etc. They have to operate under an number of constraints so there is no clear solution they could share with me at this point but I'm confident they will work something out.

...

The intent of a business is to make more money then expenses as much as possible.  If you think any business has some sort of different charity intent then I got a bridge to sell you.  It might even hash or some thing like that.  There is no solution to their desire to make more money.  They advertised MPP to collect more sales.  It had zero to do with wanting customers to succeed because any ASIC seller wants to sell their ASIC for the maximum amount possible.  Your visit reminds me of in-person accounts of visits to Enron.  Lots of fluff.

Any company that consistently screws their customers will not be a company very long (unless you are a monopoly or BFL apparently).

While I agree that HF is a company and companies will try to maximize their profits I think reality is more nuanced than that. There needs to be a balance. They need to provide value. They need to build a customer base. They need to build a reputation. But thanks for your constructive feedback.

there is a simpler explanation:

Keeping customers happy DOES maximize profits in the long run.

There are no long term profits in ASIC manufacturing because there is no added value in the devices it's only what they could likely mine in their lifetime and with every device sold and every block solved that amount gets less and less.  Once we reach the point where projected lifetime miner income is close to manufacturing cost of the ASIC the only way to sell more is for exchange rate to rise at which point all ASIC makers might as well buy BTC to try to influence the rate to keep on selling devices.

[bitcoiner49er]

both hash fast and coin terra will be at least a month late, so they both fail horribly. MPP, what a crock of shit.

By this point, anyone paying attention to the ASIC space knows to expect delays and discount wildly optimistic marketing babble about earliest possible dates given only the constraints of physics.

...

Why?  If anything you'd think by this point companies would know that delivery date is one of the most important issues for miners.

This is the ONLY important issue. Mining is 100% critical on hashing start date. Whatever price you offer for a given hashrate needs to be guaranteed by a certain date. Until the ASIC companies can understand this and deliver on it, it's nothing but vapor.
ASICminer is the only company to do this consistently. Which made it easy to figure out that their prices were too high to recoup BTC spent. Props to them for having a solid product and a solid delivery plan.

[cedivad]

TSMC, the most advanced foundry in the entire world

Google disagrees. (yea, i know, but anyway, troll for troll...)

[iCEBREAKER]

TSMC, the most advanced foundry in the entire world

Google disagrees. (yea, i know, but anyway, troll for troll...)

Your trolling sucks (almost as much as Cointerra & Global Foundries), please try harder!   

  Even you can do better than Italian Google screenshots. 

Here, have some meticulously researched, highly relevant facts instead:

AMD 28nm chips are still TSMC - GlobalFoundries couldn’t make it

http://www.fudzilla.com/home/item/30102-amd-28nm-chips-are-still-tsmc

Despite a major effort on the part of GlobalFoundries to ramp up its 28nm process and make it more stable and reliable, we are hearing that the company still faces issues.

http://www.taipeitimes.com/News/biz/archives/2013/06/06/2003564057

TSMC has cornered more than 90 percent of the worldwide advanced 28nm-chip market.  In April, the company said that revenue generated from its 28nm chips for this year would be triple that of last year, driven by strong demand for smartphones and tablets.

Sep 19, 2013   TSMC still leading in 28nm chips

http://www.taipeitimes.com/News/biz/archives/2013/09/19/2003572456

TSMC will continue to dominate the global 28-nanometer chip market this year, with sales forecast to triple to US$6.33 billion and account for 78 percent of the market, IC Insights Inc said yesterday.

http://www.extremetech.com/computing/164400-globalfoundries-may-be-stealing-28nm-business-from-rival-tsmc

Back when 28nm was on the drawing board, TSMC’s decision to go with the more difficult gate-last approach was seen as a risky maneuver. The Taiwanese foundry also announced it would ramp up multiple 28nm production lines, including options for both conventional polysilicon and high-k metal gate. GlobalFoundries, in contrast, focused its efforts on gate-first design and went with high-k metal gate across the 28nm process. Between the two companies, TSMC’s bets have paid off more effectively; earlier this year the company claimed to be responsible for 90% of worldwide 28nm production.

[cedivad]

Ok, i will try harder. You know, i'm new in this sector.

[iCEBREAKER]

You know, i'm new in this sector.

It shows.   

[ImI]

Ok, i will try harder. You know, i'm new in this sector.

icybreaker surely is a good teacher in terms of trolling. just look at his bright ignore button and you know whats the situation.

[RoadStress]

Thanks for the report.

Good to hear the adults have arrived.

Please ignore all the 'nothing has changed because I'm still whining' idiots.   

The report said nothing new (actually it said just nothing) and you post this reply?