What hardware to buy with $500K. Can you get 16NM or 20NM ASICs?

[ajmcello]

I was thinking about renting some colocation cabinets and estimate a $600+ day power bill. This would probably place my rent in the $25K a month range if I decided to get into the bitcoin business and drop some serious cash on hardware.

If you were to drop $500K in hardware, what would you buy that is available or expected to be available in the next couple of months? I would think the price of bitcoin would need to be about $250 or greater to return a profit, and I believe eventually it'll work its way on up to the 4's and $500 again. But hey that's just me, and lots of folks I've talked to say bitcoin is dead. It's been flat, and growing slower in acceptance this year, but I have hope.

I've read 16NM has been in development and I think 20NM ASICs are available. Can one buy just the chips themselves and have them assembled on a PCB? Essentially, what would be the best bang for the buck to return a profit? Or is the difficulty just too great now and power consumption too much that its impossible? Somewhere the numbers must add up to return a profit, but I've had no luck in finding the hardware needed to do so. And other companies have decided to no longer sell to the public and just mine for themselves. That's why I was curious about sourcing the ASIC chips themselves. But I have a feeling it would be a lot more work to assemble a working rig if one were to drop a few hundred grand. Is cloud hashing a good bet?

Thanks in advance.

[TheRealSteve]

I'll leave the usual mine vs buy, ROI vs no ROI, cloud vs no cloud, SPT vs BM vs Avalon etc. aside and just tackle the subject's question and related bits:

Can you get 16NM or 20NM ASICs?

Not at this time.  The only 20nm publicly claimed to have been released was used in KnCMiner's Neptune - the chip which was not available for individual sale if I recall that correctly.  KnCMiner are also the only ones with a public claim to 16nm with their Solar chip, which they'd only employ in their own mining farms.  I wouldn't get too hung up on the node process anyway.

I have a feeling it would be a lot more work to assemble a working rig if one were to drop a few hundred grand

You'll need somebody well versed in electronics design (digital and analog) to actually design that PCB for you.  Even if you take an open source design (e.g. Avalon - SFARDS has also released detailed schematics and have a dev kit), there's little reason to think you could manufacture cheaper than the source.  You'd be cutting out operational overhead, mostly.

If you did want to try sourcing chips, see:
https://bitcointalk.org/index.php?topic=1002697

Here's a decent play-by-play of getting a chip with just enough information to get it working and designing board(s) around it:
https://bitcointalk.org/index.php?topic=995675

Though if you have $500k to throw around, these things wouldn't be the stumbling blocks per se, as you could talk to a manufacturer or mine operator directly and work out a deal.

[dogie]

I have a feeling it would be a lot more work to assemble a working rig if one were to drop a few hundred grand

You'll need somebody well versed in electronics design (digital and analog) to actually design that PCB for you.  Even if you take an open source design (e.g. Avalon - SFARDS has also released detailed schematics and have a dev kit), there's little reason to think you could manufacture cheaper than the source.  You'd be cutting out operational overhead, mostly.

You could turn up with $10M in cash and still be behind for the next (upcoming) generation, almost everyone company is going for it.

[notlist3d]

I think you can look over at this thread and see current gen still has some promise of more efficiency (specifically Bitmains chip used in the S5) - https://bitcointalk.org/index.php?topic=902305.0

As far as getting that next NM no one really knows what cost will be as it's not been made yet.  Although 500k is a drop in the bucket most likely.  And really even after they are made I find chances small most companies will tell costs.

[pekatete]

Since you are taliking about dropping $500K, I'd say look at setting up a mining farm powered by the bitmain S5's in a teritory with low electricity cost (now and if possible the next year or two) as opposed to cloud mining. With luck going your way and BTC value appreciating, you can realistically be looking at making your money back in a short time span.
$500K is a lot of money by most standards, in the context of chip developement, it is a drop in the proverbial ocean.

[philipma1957]

I was thinking about renting some colocation cabinets and estimate a $600+ day power bill. This would probably place my rent in the $25K a month range if I decided to get into the bitcoin business and drop some serious cash on hardware.

If you were to drop $500K in hardware, what would you buy that is available or expected to be available in the next couple of months? I would think the price of bitcoin would need to be about $250 or greater to return a profit, and I believe eventually it'll work its way on up to the 4's and $500 again. But hey that's just me, and lots of folks I've talked to say bitcoin is dead. It's been flat, and growing slower in acceptance this year, but I have hope.

I've read 16NM has been in development and I think 20NM ASICs are available. Can one buy just the chips themselves and have them assembled on a PCB? Essentially, what would be the best bang for the buck to return a profit? Or is the difficulty just too great now and power consumption too much that its impossible? Somewhere the numbers must add up to return a profit, but I've had no luck in finding the hardware needed to do so. And other companies have decided to no longer sell to the public and just mine for themselves. That's why I was curious about sourcing the ASIC chips themselves. But I have a feeling it would be a lot more work to assemble a working rig if one were to drop a few hundred grand. Is cloud hashing a good bet?

Thanks in advance.

your question is all wrong.  it needs to be where can I get safe reliable 3 to 4 cent power?

  current gear is .4 to .5 watts

s-3 gear is .8 watts

new gear even if it is .3 watts  loses out to s-3's at 3 to 4 cent power.  why is that?  it will cost 2 to 3 times the price of an s-3

play with this calculator

https://bitcoinwisdom.com/bitcoin/calculator

if you understand the math you will see why  500k worth of s-3's in a 3 to 4 cent cost zone make sense.

[Bicknellski]

Mining speculation or I WANT TO BUY XXX$ worth of miners? Really?

Wrong section.

This topic has been moved to Mining speculation will be the next move for this thread one can only hope.

[Tupsu]

 Is cloud hashing a good bet?

Thanks in advance.

Yes. Buy BTC, and go to www.hashnest.com

Do not forget   Google Auth.

Price: 388.0 USD(1.66819656฿)  Stock: 1566
Maintenance Fee:$0.001175/GHS/Day


For 500 000 USD You can buy  1288 Antminer S5

1487 TH/s      6-7 BTC per day,  and no electricity bill at the end of the month. 

[muhrohmat]

well im not too shaby on my comment but why not a vacation with some money this ant miners give hard work to calculate good profit and roi but 7 btc like 1600 dollars per day you need 1 year to roi its very high investment in my opinion

[Tupsu]

well im not too shaby on my comment but why not a vacation with some money this ant miners give hard work to calculate good profit and roi but 7 btc like 1600 dollars per day you need 1 year to roi its very high investment in my opinion

Agree. Roy 1 year.
But you can always sell your S5 GHS,
better and faster than physical miners.   The only way to lose,  if the BTC price increases, but then rising incomes as well.

It is very difficult to keep mining 1000 antminer S5

This will require a large investment in cables, power supplies and shelves. Antminer S5 is very capricious,
you need to hire employees.

[ajmcello]

My calculator @ .10 for power and 1.2PH/s put me at around 9 months, but I wasn't looking at ANT Miner's, but rather a new product that is due out soon. I'd need about 200 of them, and might boost it to 1.8PH/s.

The problem I run into is that's a lot of colocation space. Each server is 2u, 1600W and does 6TH/s. It may make more sense to host it at home, but I'd need about 500kW of power....Lol..Not sure what is needed to get that kind of set up, but hey I do have a 4 car garage with no cars it could all fit into.


I think 9 months is a reasonable return on hardware if I'm between 1.2-1.8PH.

[ajmcello]

Really yourself, buckwheat. You must really enjoy moving threads, as we're not all hip to each category. This was a mining hardware question. TBYSADH. :\

Mining speculation or I WANT TO BUY XXX$ worth of miners? Really?

Wrong section.

This topic has been moved to Mining speculation will be the next move for this thread one can only hope.

[notlist3d]

Really yourself, buckwheat. You must really enjoy moving threads, as we're not all hip to each category. This was a mining hardware question. TBYSADH. :\

Mining speculation or I WANT TO BUY XXX$ worth of miners? Really?

Wrong section.

This topic has been moved to Mining speculation will be the next move for this thread one can only hope.

I don't think he meant to be mean with his comment.  If you were really investing 500k yes you could very well make a thread about it and your updates (It might be put in mining possibly though).   

People come here asking if I buy X miners what will happen?  If I had X amount of cash and bought miners what will happen?  At least once a week these are asked if not more.  But people just don't follow through and buy the massive amount of hardware.  So this causes it to be speculation, or something that really is not going to happen.

And again none of this is meant against you or your question.  It just is common for something like this to be moved from hardware.

[ajmcello]

Well, it was mostly a hardware mining question, since I wanted to know which miner was the best bang for the buck (one that would be out or coming out soon) or even if 16 or 20NM ASICs were available. So I don't know.

I don't think he meant to be mean with his comment.  If you were really investing 500k yes you could very well make a thread about it and your updates (It might be put in mining possibly though).   

People come here asking if I buy X miners what will happen?  If I had X amount of cash and bought miners what will happen?  At least once a week these are asked if not more.  But people just don't follow through and buy the massive amount of hardware.  So this causes it to be speculation, or something that really is not going to happen.

And again none of this is meant against you or your question.  It just is common for something like this to be moved from hardware.

[SebastianJu]

I was thinking about renting some colocation cabinets and estimate a $600+ day power bill. This would probably place my rent in the $25K a month range if I decided to get into the bitcoin business and drop some serious cash on hardware.

If you were to drop $500K in hardware, what would you buy that is available or expected to be available in the next couple of months? I would think the price of bitcoin would need to be about $250 or greater to return a profit, and I believe eventually it'll work its way on up to the 4's and $500 again. But hey that's just me, and lots of folks I've talked to say bitcoin is dead. It's been flat, and growing slower in acceptance this year, but I have hope.

I've read 16NM has been in development and I think 20NM ASICs are available. Can one buy just the chips themselves and have them assembled on a PCB? Essentially, what would be the best bang for the buck to return a profit? Or is the difficulty just too great now and power consumption too much that its impossible? Somewhere the numbers must add up to return a profit, but I've had no luck in finding the hardware needed to do so. And other companies have decided to no longer sell to the public and just mine for themselves. That's why I was curious about sourcing the ASIC chips themselves. But I have a feeling it would be a lot more work to assemble a working rig if one were to drop a few hundred grand. Is cloud hashing a good bet?

Thanks in advance.

One thing you should consider before planning to host them on a normal colocation rack... miners eat way more power than normal colocation offers can provide. You need to have a specialized offer and it wont be cheap in normal data centres.

Though there are some hosting companies, like ASICSPACE, that offer hosting and they have a very low power price. You might check one of those out. The good thing is that you wont need to check your hardware all the time, they will do it for you.

[GreatNorthData]

If you're actually looking for that kind of power demand, please drop us a line. We can do much better than your 10 cent/kwh estimate.

[jonnybravo0311]

My calculator @ .10 for power and 1.2PH/s put me at around 9 months, but I wasn't looking at ANT Miner's, but rather a new product that is due out soon. I'd need about 200 of them, and might boost it to 1.8PH/s.

The problem I run into is that's a lot of colocation space. Each server is 2u, 1600W and does 6TH/s. It may make more sense to host it at home, but I'd need about 500kW of power....Lol..Not sure what is needed to get that kind of set up, but hey I do have a 4 car garage with no cars it could all fit into.


I think 9 months is a reasonable return on hardware if I'm between 1.2-1.8PH.

The miner you're describing does not exist.  I know exactly which one's specs you're quoting and it's vaporware.  500kW into your garage???  LOL.  Now that is just funny.

[Mikestang]

It may make more sense to host it at home, but I'd need about 500kW of power....Lol..Not sure what is needed to get that kind of set up, but hey I do have a 4 car garage with no cars it could all fit into.

There's no cost effective way do get 500kW to a residence.  Any power company will laugh at you, residential service lines likely wouldn't support that, and you'd be stupid to pay residential rates for a 500kW service.

You could drop $100,000USD and get your own generator:
http://www.cat.com/en_US/products/new/power-systems/electric-power-generation/diesel-generator-sets/18331379.html

But you'll never have peace or quiet, your neighbors will hate you, and it's probably not legal to have it running 24/7 in a residential zone.

[alh]

You could approach 21 Inc, and see if they will be building anything larger than a "toaster miner". Supposedly they have really efficient chips, though it's not clear if they have built a miner or what. If you throw enough money at them, they might consider something besides "A miner in every cable box". This would reduce your electrical infrastructure costs, though probably still well outside the "Do it in my garage" plan.

Is your 6TH/1600W miner the Uranus miner of some kind?

[Coin.millionair]

You can get 16nm miner or order hosting from here: http://www.yesminers.de/en?acc=a87ff679a2f3e71d9181a67b7542122c