[ANN] Spondoolies-Tech - carrier grade, data center ready mining rigs

[RoadStress]

I was in the Gavin-train for a good period of time, but recently I have come to the conclusion that I was a bit wrong and that the Core is currently on the right track. This doesn't mean that I will support them no matter what, but trying to find out if my choice is good or not I am constantly trying to find a weak spot by asking myself questions and then searching for the answer. I will apply that to your post too in order to have some constructive discussion.

Core working on the code doesn't mean only power to do whatever they want. They are also responsible and accountable if something goes wrong. If we get to a point where the blocks are always full it is their responsibility to manage that situation. If they will not be able or if they won't do anything then everyone else will simply fork the code, change the blocksize limit to whatever they agree upon as long as it is bigger than 1MB and they will do anything in their power to continue using bitcoin the same way as they did before. Remember that it only takes one single line to change the blocksize and that is something that everyone can do so we don't depend on Core to do that.
The rage will start from the normal users who will simply reduce their number of transactions and that will impact everyone. If everyone will be impacted by the reduced number of transactions then they will take measures and the easiest way is to simply fork the Core software, bump the blocksize limit with one single line of code.

I think its good to have alternative clients though that enable a way to 'vote' on changes in a real way (rather than in reddit and forums with no real polling methods). AFAIC, I want classic to succeed in creating a fork (or at least signalling >50% hashrate support as a signal to core devs that perhaps segwit alone isnt sufficient) to a simple 2MB blocksize and otherwise identical to BTC-Core (0.12 or anything that comes after)

How would voting work? Based on what? Are we able to have a fraud proof voting system?

besides a modest 30-50% space gain through segwit that takes a few months to really be noticed, i think we will start seeing a lot of full blocks and rising fees soon. at this point, fees are already moving above what spam transactions would pay, and if this continues it will start to push out microtransactions and make important transactions more costly. without a maxblocksize increase at that point, it basically forces bitcoin users into blockstream-designed sidechains under imposed economic conditions, rather than through appealing natural improvements/benefits that the blockchain cant provide (like 1min validation)

I have learned that the wheels of change from an ecosystem move slower and slower directly proportional with its size. Since bitcoin is today much bigger than it was 5 years ago it will take a bit of time for the wheels to turn when blocks will be full and the fees will start to rise. Fortunately we already have a solution to that here and now (SegWit) which also opens the door for new improvements and buys us time in order to deploy a bigger and more complex solution (like the hardfork).

[Guy Corem]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

[sjc1490]

Shameless Plug, but I have 5 SP20 for sale if anyone interested.
https://bitcointalk.org/index.php?topic=1352995.0

[klondike_bar]

I was in the Gavin-train for a good period of time, but recently I have come to the conclusion that I was a bit wrong and that the Core is currently on the right track. This doesn't mean that I will support them no matter what, but trying to find out if my choice is good or not I am constantly trying to find a weak spot by asking myself questions and then searching for the answer. I will apply that to your post too in order to have some constructive discussion.

Core working on the code doesn't mean only power to do whatever they want. They are also responsible and accountable if something goes wrong. If we get to a point where the blocks are always full it is their responsibility to manage that situation. If they will not be able or if they won't do anything then everyone else will simply fork the code, change the blocksize limit to whatever they agree upon as long as it is bigger than 1MB and they will do anything in their power to continue using bitcoin the same way as they did before. Remember that it only takes one single line to change the blocksize and that is something that everyone can do so we don't depend on Core to do that.
The rage will start from the normal users who will simply reduce their number of transactions and that will impact everyone. If everyone will be impacted by the reduced number of transactions then they will take measures and the easiest way is to simply fork the Core software, bump the blocksize limit with one single line of code.

I think its good to have alternative clients though that enable a way to 'vote' on changes in a real way (rather than in reddit and forums with no real polling methods). AFAIC, I want classic to succeed in creating a fork (or at least signalling >50% hashrate support as a signal to core devs that perhaps segwit alone isnt sufficient) to a simple 2MB blocksize and otherwise identical to BTC-Core (0.12 or anything that comes after)

How would voting work? Based on what? Are we able to have a fraud proof voting system?

besides a modest 30-50% space gain through segwit that takes a few months to really be noticed, i think we will start seeing a lot of full blocks and rising fees soon. at this point, fees are already moving above what spam transactions would pay, and if this continues it will start to push out microtransactions and make important transactions more costly. without a maxblocksize increase at that point, it basically forces bitcoin users into blockstream-designed sidechains under imposed economic conditions, rather than through appealing natural improvements/benefits that the blockchain cant provide (like 1min validation)

I have learned that the wheels of change from an ecosystem move slower and slower directly proportional with its size. Since bitcoin is today much bigger than it was 5 years ago it will take a bit of time for the wheels to turn when blocks will be full and the fees will start to rise. Fortunately we already have a solution to that here and now (SegWit) which also opens the door for new improvements and buys us time in order to deploy a bigger and more complex solution (like the hardfork).

http://gavinandresen.ninja/a-guided-tour-of-the-2mb-fork

2mb is not a simple one-line change, and the implementation used for btc/classic is about 900 lines, based largely on coding crafted for XT. A lot of that is simply testing code, but even the core change of 2mb is almost 20 lines:

The first commit is “Minimal consensus/miner changes for 2mb block size bump”, and is small – under twenty lines of new code. You can see the one-line change of MAX_BLOCK_SIZE from 1,000,000 bytes to 2,000,000 bytes; the rest of the changes are needed by miners so they know whether it is safe to produce bigger blocks.

a node can easily do the 1-line change and it will accept and relay blocks <2mb, but most other nodes will see it as invalid and not relay them, hence making mining >2mb infeasible. Thats why it makes sense to include a "trigger" so that the client only accepts <1MB until the 2MB rule change is implemented. Otherwise XT would have made a few dozen 1-2 block chainforks already (before switching back to the main longest chain). The consensus mechanism only ensures that the 2mb fork will be successful when activated 28days after 75% majority.


As for core; I like thier roadmap. segwit has benefits to offer but they may be smaller and further away then currently anticipated, in the face of a growing network. Other improvements of the code that are planned are also generally quite good, but without an actual blocksize scaling event its easy to envision a path where blockstream is dedicated to making bitcoin a high-fee settlement layer with massive segwit sideblocks that power sidechains and lightning. Its a good direction, but I disagree with limiting bitcoin in its infancy to a 1mb limit that has gone without incident for years around improved computing and networking, particularly as it makes blockstream's business model more attractive

I firmly beleive in segwit+2mb and i think theres a developing similar consensus. in fact, segwit could make 2mb blocks much simpler by removing the complex validation needs of a small-but-many-signatures transaction.

I think its good to have alternative clients though that enable a way to 'vote' on changes in a real way (rather than in reddit and forums with no real polling methods). AFAIC, I want classic to succeed in creating a fork (or at least signalling >50% hashrate support as a signal to core devs that perhaps segwit alone isnt sufficient) to a simple 2MB blocksize and otherwise identical to BTC-Core (0.12 or anything that comes after)

How would voting work? Based on what? Are we able to have a fraud proof voting system?

its quite simple, you can identify your client  (such as  classic-0.11.2.b1) and you can view node count at https://coin.dance/nodes  (IIRC they filter out nodes if they detect multiple on a single address (ie: someone pretending to run 10 nodes on one computer)). If you are a miner, you can follow BIP009 to include a modified versionbits, or you could simply include a message or code in the coinbase transaction (almost every miner out there does this already, and many even include BIP100 voting tags)

hashrate is pretty much the most secure method of voting as it requires control over bitcoin mining. Of course, only miners can vote this way.

[tutorialevideo]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

[TKeenan]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

[RoadStress]

http://gavinandresen.ninja/a-guided-tour-of-the-2mb-fork

2mb is not a simple one-line change, and the implementation used for btc/classic is about 900 lines, based largely on coding crafted for XT. A lot of that is simply testing code, but even the core change of 2mb is almost 20 lines:

The first commit is “Minimal consensus/miner changes for 2mb block size bump”, and is small – under twenty lines of new code. You can see the one-line change of MAX_BLOCK_SIZE from 1,000,000 bytes to 2,000,000 bytes; the rest of the changes are needed by miners so they know whether it is safe to produce bigger blocks.

a node can easily do the 1-line change and it will accept and relay blocks <2mb, but most other nodes will see it as invalid and not relay them, hence making mining >2mb infeasible. Thats why it makes sense to include a "trigger" so that the client only accepts <1MB until the 2MB rule change is implemented. Otherwise XT would have made a few dozen 1-2 block chainforks already (before switching back to the main longest chain). The consensus mechanism only ensures that the 2mb fork will be successful when activated 28days after 75% majority.

The rest of 899 lines are just for consensus check and for unit tests, but the actual block size change is still one line if we assume there are exactly 0 bad actors. And in my scenario where everyone gets sick of Core and simply want to fork away with >1MB blocks there are 0 bad players in the ecosystem. At least that's how I see a split would take place.

As for core; I like thier roadmap. segwit has benefits to offer but they may be smaller and further away then currently anticipated, in the face of a growing network. Other improvements of the code that are planned are also generally quite good, but without an actual blocksize scaling event its easy to envision a path where blockstream is dedicated to making bitcoin a high-fee settlement layer with massive segwit sideblocks that power sidechains and lightning. Its a good direction, but I disagree with limiting bitcoin in its infancy to a 1mb limit that has gone without incident for years around improved computing and networking, particularly as it makes blockstream's business model more attractive

I firmly beleive in segwit+2mb and i think theres a developing similar consensus. in fact, segwit could make 2mb blocks much simpler by removing the complex validation needs of a small-but-many-signatures transaction.

If we agree that the segwit brings much more benefits for the long term than a 2MB block then I think that having segwit happen as fast as possible is the correct choice. A 2MB block now would only benefit the block space, while the segwit brings us so much more!

its quite simple, you can identify your client  (such as  classic-0.11.2.b1) and you can view node count at https://coin.dance/nodes  (IIRC they filter out nodes if they detect multiple on a single address (ie: someone pretending to run 10 nodes on one computer)). If you are a miner, you can follow BIP009 to include a modified versionbits, or you could simply include a message or code in the coinbase transaction (almost every miner out there does this already, and many even include BIP100 voting tags)

hashrate is pretty much the most secure method of voting as it requires control over bitcoin mining. Of course, only miners can vote this way.

How do you fraud proof if you identify by client version?
How would votes count? How much weight would miners have? Do they have 1 vote per entity?
There are so many questions that I could raise about this voting system that we can drag this for a very long time. I don't think that you could correctly weight every stakeholder in the ecosystem in order to have a voting mechanism in place. I think that we have so much diversity starting from the average Joe who has no idea about what the system is capable or not to the miners who operate large scale mining operations while running on a very thin margins to businesses who offer various services on the existing network. How do you make the difference between them in order to have a correct voting system?

[notlist3d]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

I really would like to know out of curosity on SP50.  If they are able to do some changing and get a chip that is closer to Bitfury's reported chip.  Or if they really will not release and consider sunk cost.

Only way I see them winning is if they can drastically beat Bitfury on chip price.  And that no one knows, or at least no one without a NDA.  For what it's worth I was hoping they would do consumer ones like SP20 again instead of SP50.  Doubt that changes but seems they would want to do something with the chips they I would guess put a decent amount into R/D.  And I guess we really don't know that either as no demo.

So many good questions, but don't think we will get anwser.  Although eventually we should it is interesting how SP has to release info compared to most privately owned firms that are tight lip'ed.

[Fatman3001]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

I really would like to know out of curosity on SP50.  If they are able to do some changing and get a chip that is closer to Bitfury's reported chip.  Or if they really will not release and consider sunk cost.

Only way I see them winning is if they can drastically beat Bitfury on chip price.  And that no one knows, or at least no one without a NDA.  For what it's worth I was hoping they would do consumer ones like SP20 again instead of SP50.  Doubt that changes but seems they would want to do something with the chips they I would guess put a decent amount into R/D.  And I guess we really don't know that either as no demo.

So many good questions, but don't think we will get anwser.  Although eventually we should it is interesting how SP has to release info compared to most privately owned firms that are tight lip'ed.

If they released SP50 with "current" specs it would be plenty competitive, but fabs need to see money.

Without some sort of pre-order it will be difficult to get this off the ground.

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

[dogie]

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

If it doesn't flat line then someone is doing something wrong.

[notlist3d]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

I really would like to know out of curosity on SP50.  If they are able to do some changing and get a chip that is closer to Bitfury's reported chip.  Or if they really will not release and consider sunk cost.

Only way I see them winning is if they can drastically beat Bitfury on chip price.  And that no one knows, or at least no one without a NDA.  For what it's worth I was hoping they would do consumer ones like SP20 again instead of SP50.  Doubt that changes but seems they would want to do something with the chips they I would guess put a decent amount into R/D.  And I guess we really don't know that either as no demo.

So many good questions, but don't think we will get anwser.  Although eventually we should it is interesting how SP has to release info compared to most privately owned firms that are tight lip'ed.

If they released SP50 with "current" specs it would be plenty competitive, but fabs need to see money.

Without some sort of pre-order it will be difficult to get this off the ground.

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

That is if they launch now or beat bitfury by a significant amount of time.  Since bitfury has shown demo of chip, and SP has not had anything but rendering publicly I'm not guessing this will happen.

I just see the big operations going to most efficiency with the "big money".   NDA's make it kinda hard to know progress as most are secret on a lot.  But if Bitfury beats SP to launch and has a better efficiency... I don't think it will be competitive.

And that is pure speculation.  We have no idea on SP process on how far they are with chip.  

[Fatman3001]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

I really would like to know out of curosity on SP50.  If they are able to do some changing and get a chip that is closer to Bitfury's reported chip.  Or if they really will not release and consider sunk cost.

Only way I see them winning is if they can drastically beat Bitfury on chip price.  And that no one knows, or at least no one without a NDA.  For what it's worth I was hoping they would do consumer ones like SP20 again instead of SP50.  Doubt that changes but seems they would want to do something with the chips they I would guess put a decent amount into R/D.  And I guess we really don't know that either as no demo.

So many good questions, but don't think we will get anwser.  Although eventually we should it is interesting how SP has to release info compared to most privately owned firms that are tight lip'ed.

If they released SP50 with "current" specs it would be plenty competitive, but fabs need to see money.

Without some sort of pre-order it will be difficult to get this off the ground.

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

That is if they launch now or beat bitfury by a significant amount of time.  Since bitfury has shown demo of chip, and SP has not had anything but rendering publicly I'm not guessing this will happen.

I just see the big operations going to most efficiency with the "big money".   NDA's make it kinda hard to know progress as most are secret on a lot.  But if Bitfury beats SP to launch and has a better efficiency... I don't think it will be competitive.

And that is pure speculation.  We have no idea on SP process on how far they are with chip.  

But Bitfury doesn't sell any of their gear. They just maintain higher profits. They're not going to trigger another 51% attack fear wave. I hope.

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

If it doesn't flat line then someone is doing something wrong.

The major uncertainty is BTC(XBT) price development.

[notlist3d]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

I really would like to know out of curosity on SP50.  If they are able to do some changing and get a chip that is closer to Bitfury's reported chip.  Or if they really will not release and consider sunk cost.

Only way I see them winning is if they can drastically beat Bitfury on chip price.  And that no one knows, or at least no one without a NDA.  For what it's worth I was hoping they would do consumer ones like SP20 again instead of SP50.  Doubt that changes but seems they would want to do something with the chips they I would guess put a decent amount into R/D.  And I guess we really don't know that either as no demo.

So many good questions, but don't think we will get anwser.  Although eventually we should it is interesting how SP has to release info compared to most privately owned firms that are tight lip'ed.

If they released SP50 with "current" specs it would be plenty competitive, but fabs need to see money.

Without some sort of pre-order it will be difficult to get this off the ground.

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

That is if they launch now or beat bitfury by a significant amount of time.  Since bitfury has shown demo of chip, and SP has not had anything but rendering publicly I'm not guessing this will happen.

I just see the big operations going to most efficiency with the "big money".   NDA's make it kinda hard to know progress as most are secret on a lot.  But if Bitfury beats SP to launch and has a better efficiency... I don't think it will be competitive.

And that is pure speculation.  We have no idea on SP process on how far they are with chip.  

But Bitfury doesn't sell any of their gear. They just maintain higher profits. They're not going to trigger another 51% attack fear wave. I hope.

In past they have but you have to have substantial order.  Home/hobby miners  are not their customers.   But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0  .  They also sold Bit-X 1 PH long ago - https://bitcointalk.org/index.php?topic=896350.0

In past they have sold to big commercial operations.  So not sure where you got info they don't sell at all.  SP I see as going for same customers as the SP50 was geared twords big operations.  But again still lots of questions few answers.

[Fatman3001]

Congratulations to Nimrod Lehavi and Simplex team

http://www.coindesk.com/israeli-bitcoin-startup-raises-7m-in-series-a-round/

Buying Bitcoins will become much easier. Awesome company.

Any news on the sp-50?

100% dead.  Ain't never coming.  Spondoolies and Guy Corem to get sued by creditors soon.

I really would like to know out of curosity on SP50.  If they are able to do some changing and get a chip that is closer to Bitfury's reported chip.  Or if they really will not release and consider sunk cost.

Only way I see them winning is if they can drastically beat Bitfury on chip price.  And that no one knows, or at least no one without a NDA.  For what it's worth I was hoping they would do consumer ones like SP20 again instead of SP50.  Doubt that changes but seems they would want to do something with the chips they I would guess put a decent amount into R/D.  And I guess we really don't know that either as no demo.

So many good questions, but don't think we will get anwser.  Although eventually we should it is interesting how SP has to release info compared to most privately owned firms that are tight lip'ed.

If they released SP50 with "current" specs it would be plenty competitive, but fabs need to see money.

Without some sort of pre-order it will be difficult to get this off the ground.

One major worry for anyone selling HW now is that there is a real chance that mining will flatline for a couple of years.

That is if they launch now or beat bitfury by a significant amount of time.  Since bitfury has shown demo of chip, and SP has not had anything but rendering publicly I'm not guessing this will happen.

I just see the big operations going to most efficiency with the "big money".   NDA's make it kinda hard to know progress as most are secret on a lot.  But if Bitfury beats SP to launch and has a better efficiency... I don't think it will be competitive.

And that is pure speculation.  We have no idea on SP process on how far they are with chip.  

But Bitfury doesn't sell any of their gear. They just maintain higher profits. They're not going to trigger another 51% attack fear wave. I hope.

In past they have but you have to have substantial order.  Home/hobby miners  are not their customers.   But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0  .  They also sold Bit-X 1 PH long ago - https://bitcointalk.org/index.php?topic=896350.0

In past they have sold to big commercial operations.  So not sure where you got info they don't sell at all.  SP I see as going for same customers as the SP50 was geared twords big operations.  But again still lots of questions few answers.

I sent them an e-mail. They don't sell hw, but they sell "hash capacity" starting at 1PH.

[Exoskeleton]

I sent them an e-mail. They don't sell hw, but they sell "hash capacity" starting at 1PH.

"Cloud mining" only?! And 1PH min orders, please.

[notlist3d]

In past they have but you have to have substantial order.  Home/hobby miners  are not their customers.   But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0  .  They also sold Bit-X 1 PH long ago - https://bitcointalk.org/index.php?topic=896350.0

In past they have sold to big commercial operations.  So not sure where you got info they don't sell at all.  SP I see as going for same customers as the SP50 was geared twords big operations.  But again still lots of questions few answers.

I sent them an e-mail. They don't sell hw, but they sell "hash capacity" starting at 1PH.

If you bring enough capital... they do sell .  You can see in past big operations they sold to:  But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0

Were talking about a lot of money, and NDA's.  I think you talk to them with substantial capital selling can happen, at least it has in past. But just emailing them for buying no you will not get far at all.

[Fatman3001]

In past they have but you have to have substantial order.  Home/hobby miners  are not their customers.   But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0  .  They also sold Bit-X 1 PH long ago - https://bitcointalk.org/index.php?topic=896350.0

In past they have sold to big commercial operations.  So not sure where you got info they don't sell at all.  SP I see as going for same customers as the SP50 was geared twords big operations.  But again still lots of questions few answers.

I sent them an e-mail. They don't sell hw, but they sell "hash capacity" starting at 1PH.

If you bring enough capital... they do sell .  You can see in past big operations they sold to:  But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0

Were talking about a lot of money, and NDA's.  I think you talk to them with substantial capital selling can happen, at least it has in past. But just emailing them for buying no you will not get far at all.

One link where someone is selling old gear and another where Bitfury explicitly says they're not selling HW. What am I missing? The Bitfury light bulbs?

[mavericklm]

For the right $$$ they sell anything!

The question is if we, small miners, will be able to buy the new miners with this new asic!

[notlist3d]

For the right $$$ they sell anything!

The question is if we, small miners, will be able to buy the new miners with this new asic!

I would agree from past actions from them.  Some random person emailing them is not going to get the options a  big investor does when they show up with 6 figure backing.   So even if they say they only sell cloudmininng only bitfury in email.... show up with big backing in a real meeting in past you can buy miners.  A email does not equal much compared to past history.

And its getting away from the question I had which is what SP does with their gear.  I think the SP50 is beat by bitfury from info so far, and it could be wrong.   But if they only planned to sell to big customers and are beat before it's out... what do they do?  Release for cheaper? Forget it and lose R/D money?  My hope is they sell some home/hobby gear.  But I doubt that happens.

[punin]

In past they have but you have to have substantial order.  Home/hobby miners  are not their customers.   But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0  .  They also sold Bit-X 1 PH long ago - https://bitcointalk.org/index.php?topic=896350.0

In past they have sold to big commercial operations.  So not sure where you got info they don't sell at all.  SP I see as going for same customers as the SP50 was geared twords big operations.  But again still lots of questions few answers.

I sent them an e-mail. They don't sell hw, but they sell "hash capacity" starting at 1PH.

If you bring enough capital... they do sell .  You can see in past big operations they sold to:  But look at like this thread: https://bitcointalk.org/index.php?topic=1167388.0

Were talking about a lot of money, and NDA's.  I think you talk to them with substantial capital selling can happen, at least it has in past. But just emailing them for buying no you will not get far at all.

One link where someone is selling old gear and another where Bitfury explicitly says they're not selling HW. What am I missing? The Bitfury light bulbs?

https://bitcointalk.org/index.php?topic=1291890
http://bitfury.com/products