Topic: Where's the catch with this? (Read 814 times)

Yea they tend to spend more on electric power and cooling on asic farm.  Regular server farm can be very expensive though to.  Server farms a lot have to have a hot or cold disaster recovery site... so two sites for most important things.   And the fire suppression in some of these buildings is just amazing.  Asic fire supression... not so much.

So IBM I would say focuses more on quality of service, as the industry is pointed at this.   Some asic farms are very very simple as far as info-structure.  There are a few that are pretty amazing though to that they spent a ton on with asics but for most part I think asic farms remain simpler to keep costs low.

I completely forgot about the heat output on ASICs. Servers run like 10 to 30 C less than ASICs.
Anyways having a 50 to 80 C chip could be bad if you are thinking long term. If anything burns out you either want to consider replacing, or just hashing at a lower speed.

those fee are not that cheap, even assuming 0.05 per hour and all s7 antminer you have 40 of them that consume around $20k per year

While it's possible I just missed it, "who" is making the offer? I ask primarily because there is no guarantee that they will exist in a year. They go *POOF* and you have essentially no way to collect on the guarantee.

How are you supposed to make the initial payment?

Offers that are "too good to be true", usually aren't real (i.e. a scam).

They would be a completely different type of data center.  Your talking about server farms where AC is what's cooling them down.

Asic data centers are massivly different in cooling.  They just have a lot more heat exhaust with so many watt's being used.  Some do fan's and try pumping new air in and old air out with shear force of CFM's.   Whall best chances are the environments that allow evaporation cooling (does not work in all places).  So it really is too completely different things.

I was talking with an IBM employee today about data center costs. It would be roughly $1,000 per sq ft to do a raised floor data center. So if you by chance have some capital. That's a rough estimate.

If you were buying this much chances are you would want hardware and buy new.  With a order that big it's likely there is a discount as they can make less per since big amount.  It does take bulk to get discount on most but for most 75k would do it I think. (Two I'm not sure is Bitfury might take more for them to even sell to you, and SP50 once release I expect to be quite high and bulk buyers).

But you would want either your own hosting center (which is hard to start) or go with someone else already doing it which with that amount again chances are hosting will give you a discount with 200T worth of gear.

Awesome - thank you, very informative!

The catch is difficulty increases, maintenance fees, both of which will erode your net revenue until your contract is at a "break-even", in which it ceases to exist. Whether this happens before or after you have re-couped your initial investment is the big question.  As per my calculations, 200 TH/s will earn you up to $90,000 in revenue in 1 year before fees. Not sure if they have been updated, but last I checked they were $0.0015 / GHs per day, or $300 per day for 200 TH/s.  Also the block reward halving will happen mid summer next year, which will cut your revenue in half.  At that point, I don't imagine their contracts will stiil be profitable if fee schedule stays the same.

Use this: https://bitcoinwisdom.com/bitcoin/calculator calculator which takes difficulty increases into account to see the real numbers, and you will see that there is no catch, but neither is there profit.

Difficulty. That's the catch. Your 200 Th/s might be good for today, but as more people begin mining and Bitcoin miners become more powerful, the difficulty in solving blocks will go up. That means your 200 TH/s will not produce the same amount of Bitcoin as it did when you paid for it. Diminishing returns over time is the fate every Bitcoin miner faces.

You calculated 180k per year, but that assumes the difficulty stays the same as it is today. That will not happen. In general the difficulty will only go up. There may be times when it drops briefly but those periods are never sustained for very long. Difficulty adjustments BTW occur every 2 weeks.

Here is a chart that shows difficulty increase over the last 9 months. As you can see, the difficulty is almost double what it was a year ago.

So the difference between buying hash power and buying physical miners, is that you will be able to resell your physical miners when the difficulty goes up. Buying hash power is fine but your betting on that level of hash power staying relevant forever (which it won't).  If you have a physical miner, you can sell it and recoup some of the costs. That can then be put into buying a new more powerful miner. You don't have that option when you only buy hash power. Your only option is to buy more hash straight out of pocket. Reselling a physical miner essentially gives you a discount when you buy a new and improved one. But then again you also don't have to worry about things like electricity or finding a place to run your miners when you only buy hash power.

$180,000 is if the current difficulty stays the same, it won't it will go up.

Next year the reward for finding a block drops from 25 to 12.5 BTC, and the fees you mention will be more than you think.

I've been researching mining, and it seems far easier and cheaper (albeit less fun) to just purchase hashing power from a provider like genesis-mining.com.

You can buy 200 TH/s for life for about $75,000, and with this power you'd make an estimated $180,000 in a year (before fees).

I know turning that kind of profit can't be that easy

So what's the catch with this?

Thanks!