[aqme28]

>Solis bought a Bitmain Antminer S9 on eBay for $2,400. >As of this weekend, Solis has mined 1.01 BTC, worth a little more than $2,584 in his bitcoin wallet. >Solis said: >"When I bought the miner, the price per bitcoin was around $1,200. I thought I’d break even in one year, but actually it’s been about five months."

So he would have been much better off just spending that 1,200 on Bitcoin rather than on a mining rig.

[danmaz74]

Even more importantly, electricity costs weren't accounted for, while they should be pretty relevant.

[Analemma_]

According to its Amazon listing, the Antminer S9 consumes 1247 watts at the wall. Assuming he ran it 24/7, over a year he would consume about 11,000 kWh, or about $1,300 at the average U.S. electricity price of $0.12/kWh. Without knowing how much he was paying for power of course we can't know exactly how much he spent, but this certainly doesn't seem like a slam-dunk win, compared to just buying Bitcoins outright.

And indeed, it couldn't possibly. The EMH almost guarantees that hobby mining can't be profitable once the coin is popular enough.

[Nursie]

Incidentally, 11K kWh is roughly the average energy consumption of a US household.

So given his current return, of 1.01 bitcoins, that means each 2.4 bitcoins mined use around the electricity that a US household does annually.

And people wonder why there are those of us who are concerned about the environmental impact...

[Valodim]

11MWh average, 0.12 cents/kWh?! Power costs more than twice as much in Europe, and the average household uses half as much. No wonder you guys (in a twisted democratic sense of the word) aren't too fond of climate treaties.

[Analemma_]

I know you're sort of joking, but that doesn't really explain it: the places where power is cheapest in the U.S. are the places where the most renewables are used. Here in Pacific Northwest power is closer to 8 cents/kWh because it's mostly hydro; Texas is cheaper than in the surrounding South because it has more wind power, etc.

[yuhong]

In this case, the server room location is Allen, Texas, likely paying commercial rates.

[hanklazard]

Good point. This sort of waste is one of the big reasons for the (hopefully soon) shift to PoS rather than the current PoW.

[yuhong]

I think commercial rates are generally lower than residential rates, right?

[bigtunacan]

Is anyone tracking the electric costs for mining? I would love to see the stats on that. What about using something like solar to power this? Then you would have your upfront costs for the solar panels, but not have recurring electric costs?

[mabbo]

The source of the power is irrelevant: you could be selling that electricity to the grid, or offsetting the cost of your home's usage. Buying a solar panel may eliminate the middle man, but there's still lost opportunity cost.

[bigtunacan]

The source of power may be irrelevant to you, but not to me. Where I live net metering is capped at 30% of my utility costs. So if I produce more than 30% of what I use the utility company sucks up the excess that I produce and keeps it for free.

[danmaz74]

Well at least they don't charge you negative prices, as instead happens to some utilities...

[mabbo]

Valid point!

[flashdance]

I used to track where large bitcoin mines were.

Mostly they're concentrated around cheap hydroelectricity, with major mines in Labrador (2c/kwh), Central China (varies, but <4c/kwh), and Georgia (unknown, bitfury likely has a special deal with the local government). A few exceptions are Northern China (coal grid @ ~3c/kwh) and Iceland (geothermal @ maybe 5c/kwh, this isn't too clear either).

Solar panels are not effective for bitcoin mining, because hardware needs to run 24/7 (a shitload of batteries are needed) and the upfront cost is high.

[mparr4]

> hardware needs to run 24/7

Why?

[simias]

It doesn't need to but given that the difficulty is ever increasing if you only mine, say, 1/3rd of the time you're at a disadvantage compared to 24/7 miners.

Now if the price of solar electricity is completely free and bitcoin doesn't crash then mathematically there'll come a point when you'll break even but it might take a long while as the mining rewards keep dwindling as difficulty increases while you spend most of your time not mining. Eventually your miner hardware or solar panels will need maintenance and that will increase your costs.

You'd have to do the math to see how long it'll take you to break even, but something tells me that it won't be nearly as competitive as those server farms using cheap hydroelectricity.

[cft]

Because miners are by far most expensive than electricity, and if you mine at 50% the effective miner's cost is 2x

[epmaybe]

I think they mean in order to actually make a profit it needs to be running nearly endlessly.

[francisofascii]

It is very cost effective by those who don't pay for electricity directly. Plenty of older condo and apartment complexes do not meter individual customers. The electricity costs are a component of the condo fees. University housing is an example.

[acjohnson55]

Externalize the costs; internalize the profits. That's the name of the game!

[projectramo]

This is the #1 factor that pro miners look at. They pay between 0-3 c/kwh

The problem is that you have the difficulty going up, and the capital costs.

Track those three curves and you have a good sense of the window for profit.

[Analemma_]

The biggest mining operations are Chinese and Russian farms getting their power through questionably legal arrangements, so those statistics probably don't exist and wouldn't have much relevance if they did.

[flashdance]

You can't bribe your way into 30 megawatts of power.

The biggest mining operations have legal partnerships with local and national governments.[1] As well, Russia isn't a major place for bitcoin mining: instead, mining is mostly concentrated in Georgia (the country, not the state) and China.

[1] http://ifact.ge/en/bitfury/

[JumpCrisscross]

> he would have been much better off just spending that 1,200 on Bitcoin rather than on a mining rig

TL; DR Why economists don't like deflation.

[chx]

This is hindsight. He would have been best off staying the fuck away. As I mentioned a billion times cryptocurrencies (which are not currencies either) have zero legal real world use cases, currently the two uses are illegal money transfers out of China and hype. It's the worst sort of gambling: you do not even know the odds.

[jchw]

>zero legal real world use cases

I pay my VPN using bitcoin. Also have a GPU bought with bitcoin. Just saying.

[astrodust]

Yeah, and people have bought pizza, too: http://uk.businessinsider.com/bitcoin-pizza-day-passes-2000-...

[ruste]

Neither does a beautiful hand-crafted glass bong, but there are many people willing to spend thousands of dollars on them and those dollars a perfectly legally useful.

[CyberDildonics]

People keep saying things like this but reality does not support this is any way. There are many thousands of places that take cryptocurrency.

[stale2002]

Crytocurrencies transactions are significantly cheaper than wire transfers.

That's a real world use case.

[sireat]

Cryptocurrencies are not always cheaper. They used to be but not anymore.

In Europe, I split a 0.36 Euro SEPA fee for a 1000 Euro transaction between two different banks in two different countries.

Bitcoin would have cost me $3-5 if not more.

[jerrylives]

>zero legal real world use cases

Except for paying people for services and goods with bitcoin? There are plenty of legal venues to use bitcoin.

>It's the worst sort of gambling: you do not even know the odds.

So...like all investing? lol. Like seriously dude what the fuck do you know? Can you please take your blowhard bullshit somewhere else?

[paulpauper]

The major problem with buying mining rigs is that they may be pre-mined. Often the developers will use the rigs for mining and then when the rigs stop becoming profitable (assuming the bitcoin price is stable), sell them as 'new'.

[arcticfox]

This doesn't make any sense at all as a major problem. Used vs. new is basically irrelevant as long as no one is lying about the hashrate. As long as the sale is immediate there's basically no problem at all there.

The actual problem, especially in the past, was that you could pay for the hardware but not actually obtain it for months, by which point it was no longer profitable. Presumably, the manufacturers were using the hardware themselves during that period.

[SippinLean]

It was certainly a major problem, if you bought a new car but the dealer gave you one with 50k miles on it would you shrug it off if the gas mileage was still the same? At a basic level it's fraud.

There are practical concerns as well, these machines are usually pretty poor quality and the immense heat + poor cooling typically means the chips will burn out one-by-one over time. The PSUs are especially poor and will burn out and need replacing as well. Getting a used ASIC means these will be problems sooner.

[saurik]

This analogy makes it sound like there were a certain number of Bitcoins inside the computer and they are all already discovered... why does it matter if a computer is used or not: all that should matter is how fast it is (and that would be difficult to lie about as you would notice immediately upon receiving it if it wasn't as fast a second claimed), and the buyer should know how fast of a computer they are willing to buy at a given price.

[astrodust]

You sure do get a lot of suspiciously dusty gear from some of these vendors.