I am not one to continually update books, but, since this space is very dynamic I reached out to one of the experts in China, “Bob,” who helped me with chapter 5.  Bob works with large Bitcoin mining farms, specifically helping them source energy but also helps procure mining equipment too.

Here are some of his comments:

I think the value per gigahash is difficult to define. Chip Maker. Assembly. Operator. Electricity Provider. Pool.  Those are five distinct industries. I believe the hashing contracts that you hear about includes one year of electricity and all of the above.  Whereas most mining businesses quote mining solutions to operators who need to buy the machines.  I think it’s not easy to lay out the relationship to a new reader, but at the same time you have to avoid confusing different segments of the mining industries who are normally just concerned with their one segment. A tough middle road.

For instance, earlier today, Guy Corem, CEO of Spondoolies Tech, an Israeli-based Bitcoin mining manufacturer posted several comments on Bitcoin Talk:

We do have large customers. The real threat is that one self-mining ASIC provider will be able to produce a killer ASIC.
Everyone of them are trying. Buying from them help their deployment and R&D efforts. As I wrote short-term gain, scarifying the long-term.

Continuing Corem states:

Spondoolies-Tech have clear technological advantage now. We have much more impressive 3rd and 4th gen under development.
I think that if we wanted to raise funds for mega farms, BitFury style, we could have done that.
KNC is doing that right now for example. So are other players.
BitFury doesn’t need to do that, but fortunately, they are very delayed with their next gens.

I don’t think it’s good for the Bitcoin ecosystem at all.

KNC has proven track record of treating the ecosystem as a retirement fund.
CoinTerra switched to self mining (and selling unprofitable cloud contracts) and exchanges to fiat heavily on the open exchanges, exactly like KNC.
BitFury at least has the sense not to exchange in the exchanges (selling at 5% markup for high net-worth individuals who wants newly mined coins), but BitFury have unclear past and are partly responsible for the CEX.IO fiasco. I don’t think they can be trusted not to harm the ecosystem.

And another notable comment:

Not desperate at all. Just analysed and know all the known competition cost.

The trend is clear if you remove BitFury 2 last DCs (June and August): http://bitcoin.sipa.be/growth.png
At current BTC price and their machine cost, they’re almost loosing money at Georgia and Iceland.

btw: They didn’t need the $20M they raised. I can’t elaborate more.

BitmainTech margins are very, very low.

ASICMiner is selling almost at cost to recoup their $6M wafers gen3 investment.

Cointerra is also probably loosing money on their 1st gen. Their 2nd gen won’t arrive until Q2

Should I continue ?

Please quote this message one month from now. Let see how the growth graph will look.

For comparison, I reached out to Bob and asked him his thoughts on what Corem was stating.  According to Bob, in terms of Chinese manufacturers switching to self-mining:

They’ve already been doing that for the past two months. But not mining themselves, they’re all into coop mode now. The manufacturers issue the machines. The site operators invest on the sites. They split the income between them.

Spoondoolies is no bitmain tech nor bitfury. But those guys have a good chance of being no.2 if either of these guys slip up. Their chip design is highly competitive, they just need a strong hook up to a manufacturing partner in shenzhen.

A lot of the claims in the thread are exagerrated. It is fair play though I think. The fud points I see there are the centralization claims and the mining is more profitable than selling. Retail consistently overestimates the production value which means that as long as your pricing strategy is good, your machines often never ROI. Would you rather sell something that doesn’t ROI or have your capital locked up for the next 6 months?

Spoondoolies is a rare case in that they still have a lot of clout to rely primarily on pre-selling their gear, which means their product might still be quite profitable for the buyer when it comes out. So to them it seems like it is more profitable to self mine, but the fact of the matter remains that without cheap pre-orders, they don’t have the money to start a production run.

And for perspective, to give you an idea of how important this space is in China, be sure to see this new summary, “The 2nd Private Meeting of Bitcoin Mining Industry in China” from Bitell.  It will be interesting to see how many manufacturers and farms can stay afloat while market price hover above cost of production for more than just a few months.