Considering a bitcoin mining hardware purchase? Make a more informed decision with our Bitcoin Mining Dashboard.
Butterfly Labs (BFL) recently announced a new $4,600 bitcoin mining product that operates at 600 GH/s. The new miner is called the Monarch and will be the same form factor as a standard graphics card. Their second generation ASIC chip will use a 28nm process compared to the 65nm of their first generation. Application specific integrated circuits (ASIC) are specialized chips that perform bitcoin mining calculations. ASIC nodes are measured in nanometers (nm) and smaller nodes allow for faster and more power efficient designs. The new 28nm chips will allow for significant improvement in both regards.
Prior to this announcement, BFL’s largest mining rig ran at 500 GH/s and cost $22,484. It required over 100 chips and an enclosure of almost two cubic feet. The new 600 GH/s device will be the first ASIC miner to take the form factor of a standard graphics card. It has both a USB 2.0 as well as PCIe connector allowing for either standalone or server-mounted operation. For comparison, the fastest graphics card miner, the AMD 7990, costs $900 and runs at 1.2 GH/s, 0.2% as fast as the projected speed of the BFL Monarch. Additionally, BFL has announced they will have a hosting option available, though the pricing for this service is yet to be announced.
They are joining the ranks of at least three other mining companies that have announced 28nm ASIC designs, though none have demonstrated working prototypes yet.
Butterfly Labs does not release official lists of pre-orders. However, a community-reported list estimates over $10M in pre-orders, many of which have yet to be shipped. Despite being ordered over a year ago, a majority of the products waiting to be shipped are the larger “Singles” and “Mini Rigs” amounting to over 400 TH/s. BFL claims all existing pre-orders will be shipped by the end of September, which would almost double the current network speed of 450 TH/s with BFL’s shipments alone.
BFL has offered to convert existing customer’s pre-orders to orders for their 2nd generation chips after a 10% fee. Converted orders will lose their spot in line for the first generation device. Officially, BFL appears to discourage this option on their website,
“We do recommend you keep your remaining 65nm orders and take delivery because you’ll get your mining gear much sooner…. It’s also worth pointing out that FPGA era products were still very valuable well into ASIC generation rollout, and the relative power difference between 65nm and 28nm is much much less. ”
Butterfly Lab’s original FPGA singles ran around 800 MH/s, while the first generation singles run at 25 GH/s, a 33x increase. The Monarch’s estimated 600 GH/s is a 24x increase over the single, so it does not appear the second generation increase will be insignificant. Given the small size of the Monarch it would appear that this will be the lower end of their 2nd generation ASIC product line.
Butterfly Labs went through months of revisions addressing power draw issues, which ultimately resulted in them having to use more chips and larger cases to meet their original specifications, as well as missing their power goal by 300%. Although they may have learned important lessons during that experience, the PCIe cards will require a new design so delays have a strong possibility of occurring again.
Their website claims their design is in the final stages of development and will be sending wafers into production shortly.
- We’re now we’re in at the final stage of development (Tapeout) and are sending wafers into production at the foundry in the next few weeks.
- Foundry production takes 10 weeks.
- Bumping, Slicing & BGA packaging takes approximately 2 weeks.
- Initial shipments begin and ramp up to full capacity over the following 3 weeks.
Interestingly, they claim to have been working on chip development for 6 months now, meaning there was significant overlap between the period they began 28nm design and were working fixing delays in their 65nm product.
Customers now face the difficult challenge of deciding whether to roll over their current orders or to keep their spot in line. A Monarch delivered in December will be break-even at around a 75% monthly network increase through July. If shipments are pushed back to January, anything over a sustained monthly increase of 59% stops being profitable. Considering that the 500 GH/s rigs have not shipped yet, they seem the most likely candidates for conversion, though realistic expectations for shipping dates will be crucial to any decision.