Topic: ASICMINER: Entering the Future of ASIC Mining by Inventing It - page 1293. (Read 3917029 times)

Since you're discussing the pro and cons of a mining operation I think I'll throw in my recent contribution to the board. Generally economy of scale does apply - but there are other factors you need to account for, which are not directly related to economic factors. Talking about the benefits of self-mining vs. exclusively selling hardware should be obvious for all major players involved in the production of ASICs.

1) Mining operation
0.06-0.07$/kWh certainly sounds competitive on a residential level. Given that the long-term plan incorporates a sustained mining operation, the option to take advantage of excess power is attractive. Power at peak hours tends to be more expensive, than power during low demand. This special contract may give you an edge over other miners - especially since mining is an opportunistic activity. That means that you need to cycle your mining operation depending on the power rates which unfortunately reduces your average hashing capacity per chip.
That said, at the current price/difficulty ratio I don't think power bills are the limiting factor. You'd need to throw anything you got at the network until the profit margin hits 10% (generated BTC - power costs). Thus power costs may be an issue in a year from now, when all GPU miners are driven out.

As with the target size of the mining operation I am not convinced a huge mining operation is beneficial in the long run (short term yes), because the bigger the mining operation the larger the overhead costs (area, cooling, staff) and you also make yourself a target for cyber attacks. Also, the bigger the mining operation, the more you cut into the remaining profit margin of your potential customers, forcing you do reduce the selling price of your hardware. A model where you mainly act as a supplier of a few larger mining operations or retailers is more reasonable. For that you'd need to work on customer relationships with these operators and maybe provide some volume discounts and premium support.

In any event, I still think a small mining operation is key to design optimization and gives you an edge in any further development of your chips, because it shows you potential bottlenecks in overall performance.

Certainly there have to be economies of scale for GPUs when it comes to buying electricity in bulk and cooling them in bulk.

Idiot forum posters....lol....who write without thinking at all....lol....basing themselves on unexamined consensus of idiots....lol....as if that makes it reality.....lol....and claim those who think don't because they come to different results....lol.

You *thinking* the food you eat off the hands of your parents and the roof you use through the good offices of the same parents (who quite possibly in turn leech it off some welfare state or other) is free does not make it so. Random joe blow "miner"'s inability to account for his costs correctly does not make those costs zero.

Simple scenario A: you get a rig, place it somewhere in your parent's house, there's a break-in, some douches trash the place. Your rig is destroyed. Total loss. Odds of this happening? Say 5% a year.

Simple scenario B: you hold a party, someone accidentally pukes/drops a full keg/whatever on your rig. Rig is ruined. Total loss. Odds of this happening? Say 10% a year.

Simple scenario etC: there's a fire. Rig is destroyed. Total loss. Odds of this happening? Say 3% a year.

Yes, I know that events which you haven't personally witnessed can never happen, this is why you can never get a girlfriend, cause it's never happened to you in the past and consequently it will never happen to you in the future. Nevertheless, your costs, your costs are 18% a year in those examples whether you like it or not, whether you agree with them being that or not, whether you know about it or not.

Someone hosting a million of the damned things can reduce costs under 1%, and likely under 0.01% and beat you out of the market predictably enough. It won't happen in practice because you'll beat yourself out of the market first (for instance by getting a girlfriend, or by talking forum shit to someone irl and getting pounded into a pulp, or through whatever other means). But just because you're too stupid & irrelevant to matter in the discussion does not mean the invisible hand ain't there.


You are happily confusing the thing itself with one bad implementation thereof. The logical equivalent of your post is "Oh, computers my ass! I bought a Microsoft one once and within the week it kept opening flashing tabs all the time."


In fact "business plans" which confuse opportunity with leeching are the best predictors of "business" failure. Review the forum, we have ~50 or so documented cases of failure, you'll be amazed.


Exactly so.


The argument was not brought about GPUs specifically because it does not work for GPUs, so nice red herring you got going there. Are you aware, by the way, and doing this deliberately, or does your brain just not know how to work properly? Honest question.

ASICs are economy of scale manifesting itself in the mining market. They're purpose-designed GPUs practically speaking.

Let's forget about ASICs for a moment, and consider off-the-shelf GPUs or FPGAs. What is the advantage of a high-volume minjng business in my city over my own in-house mining op, per dollar invested?  I build a $1300 rig and mine, versus investing the same amount in a mining company. The owner takes their cut. Where is the advantage in larger scale?
Moving on to ASICs, there is the issue of NRE cost. Clearly, a big business can invest and achieve better efficiency. A hobbist cannot even develop their own ASIC - unless we incorporate (in the old sense of the word - pool resources together to achieve something big, like building a city or a bridge). Now we can all grow our tomatoes and mine our coins and live happily. There is no reason to go big. Distributed, p2p, in numeris vires. 

There's a difference between economies of scale and efficiencies, and your personal hobbies. You enjoy growing tomatoes, and DutchBrat enjoys running mining hardware. That's great. That's on your own personal time though, not your "productive", "working" time.
Now, if you were to start growing potatos, then cabbage, and all your other vegetables in your spare time, and started taking time off work to go deer hunting, would that be an efficient use of resources?

If you like gardening and grow tomatoes on your own instead of buying them in the store, it's not that economies of scale don't apply to that market. It just means that you're willing to sacrifice your personal time to grow your own tomatoes less efficiently than a specialized producer could. It's the same with mining hardware. I would posit that the majority of people here lose money on mining and bitcoin activity if you actually factor in a reasonable wage on the time. Run a 6GH/s farm in your basement right now sucking [email protected]? You might be making $3/day. Even outside the noise in the basement, you're making $21/week. Even with free hardware, most people would be better off spending an hour of overtime at work per week than spending it managing their mining hardware. People do it because they enjoy it, but that doesn't make it an efficient use of resources.

The fact that some miners ignore costs doesn't mean that those costs don't actually exist.  And I'd hardly be taking miners' words for anything to do with business.  I've never heard a miner say that their mining investment opportunity would lose money for investors.  Does that mean they've all been profitable?  Or even that most have?

There may well be some miners who get "free" power and "free" rent who could fit in 1 rig but not 10.  But there'll be others who could fit 10 in on the same basis.

"free" is in quotes because of course the power and rent isn't actually free - SOMEONE is paying for it.  And I'm not a huge fan of "business plans" (in quotes as most around here aren't plans and have little to do with business) that include assuming power and rent are "free" - as those plans run into a heap of trouble if down the road they turn out no longer to be "free".

I grow tomatoes on my patio, they taste great. The 19th-century "economies of scale" logic would mandate me and everyone else quitting this nonsense, so a huge greenhouses can be built thousands of miles away using loans from international banks, with patented, genetically engineered tomatoes that can survive trucking and customs clearance, and a a network of supermarkets that we all drive to in SUVs whenever we need a tomato or two, and an army and a cemetary to ensure cheap oil for the whole operation. Economies of scale, my ass.

Lol, you always make me laugh when you write before you think!

I have never heard a miner say there is cost involved in giving one ASIC or any other mining device a roof over its head. Try to get free rent when you have to give 1 million mining devices a dry place to stay....

Economies of scale.... lol.....

It depends when you start selling. Of course buyers should calculate before. The thing is that other vendors will sell Asics anyway, so at some point it will make more sense for Asicminer to sell Asics then to mine with them. For example if Asiminer has to mine 6 months to get the creationprice back in and a buyer has to work 12months with such asic it would still be a good product to sell. I mean what investment brings 100% return per year? But at that point it could make more sense to Asicminer selling many Asics because the short gain is better than wait months for the same effect. And obviously friedcat and team already took that into account when creating the business plan.
Plus they plan to develop newer and better asics depending how the first one works. So it sounds to me they want to have a name as a vendor for Asics at some time. At least friedcat wrote this way like they want to create newer, better versions to sell them and that this will be the way the company will move to in future. But of course at the start mining themselves will be far better.

How have you come to this nonsense?

Read up on economy of scale. Anything that can be done in onesies can be done in millionsies for less money. Anything.

...with Christmas colors! Don't forget to mention the Christmas colors!

So you admit you are just making shit up?

Mining profitability will follow the same trend as before - down. Faster the rigs and network, faster it will happen.



Red is mining, green is selling hardware.
This is just 2 lines for illustration. No scientifically accurate data is used.

Not necessarily true:

It wouldnt be economical for 1 person to house and power 10 mln ASICs, yet it would make perfect sense for 10 mln individuals to house and power 1 ASIC each.

Hence selling instead of self mining would make perfect sense at some point in time.

If I remember correctly that's where the 50TH self mine cut off point comes from. Exactly these reasons.

At that point in time the production costs of 1 ASIC are probably next to nothing since the Mask costs have been recouped, so why not sell as many as you can for next to nothing plus a little bit :-)

The flaw in this contorted plan which you so contortedly expound upon is that the simple offer to sell rigs in that context amounts to a declaration that they aren't worth buying.

I refered to the business plan, right. It stated that asicminers business is to first mine with the hardware as long as its bringing the best income and later start to selling Asics. I guess the time when selling Asics is a better practice then simply mining wont come so fast but it will come. Question is how far this is in the future.

I think the claims to sell fast, sell to shareholders and so on shouldnt be followed. Asicminer would be stupid to sell the Asic-Cashcows for low money while they could make a fortune selfmining. It would be unfair to the rest of shareholders too that didnt get an Asic. So no, i think selling shouldnt come fast. The question is when it will come (competition and so on) and when preorders, a buying website and so on will be available. If its planned in some months, maybe because it can be seen that the competition will grow to an amount where the win for selling an Asicdevice will be higher than mining a year or whatever then it has to be thought about marketing. Making people eager to want one, prebuy, inform them about the power of the asics (at the moment the infos would be rare of course). But this would only mean more sells later. And when you can advertise for nearly no cost at your target audience its something to consider.

Of course... when the time, when mining with asics isnt so much worth than selling it comes in years only then you wouldnt need to consider. But selling asics was part of the plan from the start so it will come most probably.

Meh, people don't see past the obvious

Funny, I've re-read the quote numerous times, but I've failed to find a single instance of "if".

With #3, from what friedcat has said in this thread, this is what seems to be going on: the chips are now out of the fab, are being packaged and tested, and soon as thats done they will be shipped to friedcat or another company to be inserted onto boards (he hasn't said whos manning the solder gun on this yet, or if he did, I missed it). He hasn't said whos manufacturing the boards (this should be easier to do than with FPGAs, btw) or if they're done yet or not.

The packaging, shipping to next location to be board mounted and have the assembled product QCed, and then being shipped to wherever friedcat has chosen to set these up for mining, can all be done concurrently depending on how the companies handle it.

Also, from what I understand, soon as mining has produced enough money for a second and possibly third run of ASICs, thats going to happen too.

I think that somewhat answers your question.

An encouraging update! It means the chips at least passed the QC at the fab, and that slicing+packaging+testing is already under way. I'm hungry for more info, please return soon and fill in more blanks:
1) Recovered the ~5000 BTC from GLBSE yet? (I would like to just stop worrying about it...)
2) Did the fab really process 12 layers in 7 days, or was it simply information delay?? [Dec 05: "There are 12 layers left", Dec 12: "we are now in QC"]
3) What is the reason for deploying the second half of the chips later than the first? [Genuine question, I'm simply not smart enough that the reasons are obvious to me!]

I think I can say on behalf of all (at least many/most) stakeholders: Thanks, congrats, and keep up the great work!