Bitcoin Magazine

Show Menu

Bitcoin Developer Confirms Butterfly Labs ASIC

Bitcoin developer Luke Dashjr has confirmed that a prototypes Butterfly Labs ASIC is now hashing. The device is still far from full capacity, pushing out only 25 GH/s at a power consumption of 180 watts, but this is nevertheless the first definitive proof that Butterfly Labs is producing a legitimate product, and is not too far from finally releasing its first batch.

Butterfly Labs’ Josh Zerlan also recently provided updates on the state of Butterfly Labs’ production in an IRC channel. The core of the conversation is this:

BFL_Josh: Well guys, I had planned on updating everyone with a video of a board hashing here in KC tonight, but I haven’t been able to get that together yet, so I’m probably going to have to push it off until tomorrow. We are targeting a start of shipment next week, but I’m not quite ready to commit to that at the moment, given our past estimates. It’s imminent, though.
Lab_Rat: It hashes????
BFL_Josh: Yes, it hashes

Further down in the conversation, Zerlan provides the main reasons for the current delay. Zerlan writes: “We may miss our power targets, that’s been part of the hold up… we think there’s a problem with the power consumption and we’re trying to figure out where it’s having an issue … What’s causing even more consternation is the fact that the wafer we burned for tests runs at far less power than a second wafer we mounted on the BGA package… so it may be a wafer by wafer thing, and since we only have two datapoints, it’s hard to nail down the issue.” To many in the field, these difficulties are unsurprising; in mid-January, Avalon founder Yifu Guo wrote on the topic “recently they changed it to 1.2W, but they won’t even reach that. We ran 65nm simulations and they should be around 3W.” But to many of Butterfly Labs’ customers, who have now suffered from six months of delays, power consumption does not even matter; given that every extra day represents a lost opportunity for profit that will never come back, almost any level of power consumption is acceptable if it means that the devices will ship faster.

Butterfly Labs’ shipment has been awaited by the community for nearly ten months; the company was in fact the first to start accepting pre-orders in June 2012. Although the original shipping date was scheduled for October, the company suffered a number of delays that changed their expected shipping date first to late November, then early January, then mid-February and finally where it is today. In the meantime two other major ASIC producers, Avalon and ASICMiner, have also started hashing, and are partially responsible for raising the network hashpower from 20 TH/s to 55 TH/s over the past three months (the other contributing factor being the rapidly increasing BTC price). However, the community is still watching Butterfly Labs intently for one key reason: its potential hashpower. Although the output of Avalon and ASICMiner has been small, with Avalon’s first release being 20 TH/s and ASICMiner’s launch 12 TH/s, Butterfly Labs’ preorders altogether make up over 60 TH/s – more than is currently on the entire Bitcoin network.

For the security of the Bitcoin network, it is arguably quite fortunate that Butterfly Labs has been delayed by so much; if their machines had come any earlier, or even today, the company would have had access to more hashpower than the rest of the entire Bitcoin network put together, fundamentally compromising a key assumption of Bitcoin’s security. This threat was a major reason why Avalon founder Yifu Guo decided to go ahead with his project; “that was our main goal – we wanted to prevent this potential monopoly,” Guo wrote in a recent interview. However, it turned out that the major providers released in just the right order; Avalon started shipping their first, 20 TH/s, batch first on Jan 19, although all but two of their devices were delayed by many weeks after that date. ASICMiner started hashing on Feb 14, starting off with 2 TH/s and slowly ramping up over the past two months, and over these past two months hundreds of anonymous GPU miners turned on their hardware because of the increased mining profitability from Bitcoin’s rapidly growing price.

The rapidly rising Bitcoin price has been a massive boon to the Bitcoin mining industry in general. Since the beginning of the year, the Bitcoin price has risen from $13.3 to over $100, making mining extremely profitable for those lucky enough to have the hardware to do it. The fact that Bitcoin mining is now so heavily focused on specialized hardware compounds the advantage; while in 2010 or 2011 such price increases were quickly met with large numbers of off-the-shelf GPUs joining the network, now the most powerful miners are all specialized devices, and production takes months to compensate for the increased demand. When Jeff Garzik received Avalon’s first ASIC at the end of January, the device paid for itself in nine days, and its profitability since then has only increased. As a result, Avalon has raised the price on their third batch to 75 BTC, or $7,500, and has nevertheless sold a significant number of units. Because ASICMiner is not selling their hardware, instead selling shares and doing mining in-house, they can benefit from the increased value of their revenues directly. Notably, Butterfly Labs has not increased their prices, but even if they continue to keep their prices the same they will still benefit massively from increased sales.

In the next few months, Bitcoin network hashpower will only continue to increase. Avalon’s three shipments altogether will make up a total of 1500 units, or over 75 TH/s, and ASICMiner is planning 50 TH/s by the end of April, and 200 TH/s soon after. By the end of the year, ASICMiner’s friedcat writes on Bitcointalk, ASICMiner’s total hashpower may be as high as 1000 TH/s, and friedcat even adds that “some may say that 1,000TH/s at the end of this year is too conservative.”

Get Top Stories Weekly

We respect your email privacy

  • Marc Christiaens

    And today is APRIL 1!

  • frknnutz

    Regarding the estimated return on the Butterfly Labs MiniRig, why would they sell them? It seems to me that they would make much more money deploying the equipment for their own mining operations. They may be doing that already. Seems that they sold a dream machine that they could actually build and with the cash from the pre sales, they built their dream machines and are now quickly refunding pre orders faster than any vendor on the internet. It is not in their best interest to release any of them for sale to the public. At the current difficulty factor, the minirig would optimally bring in over $31,000 a month. Avalon just announced that batch three will be for their own mining. I suspect that the GH/s machines were never meant to be release for sale with the exeption of a token few. Mark my words: The Bitcoin economy would suffer from centralized terra hashers do to the lack of physically diverse nodes and too many transactions being centralized in these hot-rod ASIC miners.

  • Max Power

    Except that he didn’t receive it. It’s still at BFL’s facility, and the hash rate/power consumption ratio is far off from what is advertised.

    • Umbra

      Looks like close to What Avalon has been doing at a moment…

  • BarakaX

    Butterfly Labs, alone, has over 20,000 orders pending for their 60 GH/s BitForce Single SC miner. When that massive amount of processing power comes online, it’ll add 1,200 TH/s of hashpower to the network. There are thousands of orders pending for Avalon’s latest rig, which word on the street says is processing over 80 TH/s. That’s at least 500 TH/s of hashpower right there. Plus, what about others? I think we’ll be sitting around 2,000 TH/s network-wide in a year’s time. That’s *WAY* above what this article predicts.

    It is highly likely that those who have yet to receive any equipment will be badly burned on their investments over the next year. When all that processing power comes online later on this year, difficulty will increase exponentially every step of the way and you’ll be lucky to be rewarded with 25 blocks a year with one of the new rigs. If current prices stay the same or rise, then they’ll get out alive. But if they don’t…

    People will be angry and BTC may drop in price, but Bitcoin is definitely here to stay. It’s been tested again and again, and has passed every time with flying colors. This will be just another test. Too bad it’ll be a hard lesson for so many who tried to get in on the current “Bitcoin Rush” that’s underway right now.

  • Booom game

    Butterfly labs doesn’t have a working product. PCB design is ridiculous, chip existence is also questionable. Many of the first preorders will never get break even, btc wise. It has done a lot of harm to BTC community in general. Instead of quadrupling the decentralized FPGA production, globally, everybody froze and waited. Not sure why would anyone believe this guy. I could say also that I got my first ASIC 20nm, it’s hashing beautifully, but I don’t want to take pictures and confirm it. SCAM level – 95/100

    • Umbra

      There may not be 20nm ASIC at a moment. Currently, I am asking one question, when the ASIC bubble will blast. If you like to obtain more coins at this moment, then You shall get some BTC-bonds at, also get more GPUs (I think I will now need this later,but not now)

  • Lukas Tiberius Edwards

    It is a bit cheeky to say they have shipped, but saying what is clearly a test bench model is in fact the machine that belongs to a customer, in order to avoid paying money to a charity… in fact the more I think about it, the more shady it seems.

    No postage receipt, no shipping as far as I am concerned and BFL should pay up.

  • Stonedblues

    I’m pleased I got my refund from BFL 3 months ago, I’ve made more holding those refunded bitcoin than if I would be still waiting fort them to ship.

  • Bitcoinfinger

    Anyone else notice that It didn’t actually ship at all? Luke-Jr even said it didn’t after being pressed to explain how he got it so fast on Easter. Elsewhere he said he did in fact receive it, so at some point he definitely lied about it, yet he’s the only one that received one. (But didn’t.) These products are still vaporware.

    • Umbra

      If it is, Then FGPA Circult like Ztex can do the remain, just slower but easy to expand

  • Ade

    If they’d all stop trying to out hash each other, we could save a fortune on equipment and all mine with CPU again.

  • Ben Bommhardt

    Wait. This is a company that promises high-tec products, take foreign capital in form of pre-orders, shyly demonstrates a working prototype (that does not even meet the specifics) about 10 months later, still has not a single unit shipped?

    I’d say: when this company does one of the following things, its time for a reaction of the customer side:
    – hold back the mining equipment, just to use them as source of further money, sponsored by lies, held back information and using false information to keep the hope in the customers alive.
    – use the capital to drive research regarding the efficiency of bitcoing mining devices, using this technology for own purposes or selling it. Again, based on the foreign capital of clueless customers

  • Anon

    I’m still not confident that they’re not running a scam. I won’t be sure enough to buy a Jalapeno from them until I see a pic of a room full of complete devices ready to ship… and since Avalon is outside of my price range, it looks like I’ll just be keeping the GPUs running. Thankfully the difficulty hasn’t increased too much and I can still make a coin once or twice a month…

    • Umbra

      Avalon is absolutely sure as there are nodes which are Avalon powered. With the no BS,Customer commenting for sure. However the cost for Avalon ASIC is far more expensive for late comers as Avalon only accept Bitcoin right now. Here is a better way. You will use the BTC-TC for generate coins (Bond first like real world) while you use GPU to keep mining, You may need AMD 7970 (For Simple purposes like me) or FGPA circuits (For easier expansion, for low-power consumption, but requiring some skills to do) to do the job faster (with different rate of Virtual currency to adjust your step).

  • Guest

    If these things were a good investment (could mine enough to pay for their cost), THEN the company making them would have no incentive to *EVER* sell one. Better to retain all manufactured units and put them to use mining. It’s more profitable. Therefore, if you find an off the shelf mining unit, chances are it has been priced so high that you can’t make (much) off it, because if it wasn’t, the vendor would keep the unit!