Intel Develops ASIC Miner and Enters the Crypto Mining Industry

The reason Intel has decided to enter the burgeoning domain of cryptocurrency mining is obvious, as the industry blooms exponentially and settles a strong foundation with the United States now commanding the lion’s share of the world’s hash power, Intel wants a piece of the digital pie going forward.

Major brands, not solely based in the production of ASIC crypto miners, have been developing dedicated hardware for Proof-of-Work for at least two years now– Nvidia and AMD included. Nvidia’s CMP lineup of graphic cards, a collection of “reject” GA100 accelerators that did not meet quality control standards for whatever reason, probably due to insufficient dye characteristics, on their enterprise-grade offerings.

AMD followed suit when in mid-2021 they announced their own crypto-focused graphics card for the public market based on Navi-12, manufactured exclusively by OEM partner XFX.

However, both leviathans of the GPU market have limited themselves to what they do best, GPUs. Intel on the other hand has now forayed into another, albeit more saturated, direction. They’ve built chips for an ASIC (application-specific integrated circuit) miner to rival the likes of Chinese manufacturers such as Bitmain, Canaan, MicroBT and Innosilicon. Unfortunately, though, Intel’s reveal of the performance have fallen abysmally short of expectations. Intel planned to introduce a new “Ultra-Low-Voltage-Energy-Efficient Bitcoin Mining Asic” since 2018, and it seems the performance comes from that era as well.

Intel’s new ASIC miner is built with the aim of developing “the most energy-efficient computing technologies at scale.” How exactly Intel plans on doing this though is a little bit confusing and will be revealed in depth at the upcoming ISSCC conference on February 23rd. The idea Intel has is to develop specialized chips that require less power and fewer resources than traditional chips used in ASIC designs. However, it appears that although the intention is there, the final product is far from competitive in this regard.

The chip in question is dubbed the Bonanza Mine” (BMZ1) and is unique because of its small size, measuring just 7nm. Each ASIC from Intel will carry 258 of these individual mining chips to allow the ASIC to perform at 40TH/s with a power consumption of 3,600 Watts. This is a result that is not only far from desired, but most likely counterproductive. To put this in perspective, Bitmain, one of the leading ASIC manufacturers has recently announced its new next-gen model to the public, the S19+ Hydro. It is a liquid-cooled closed-loop design capable of up to 198TH/s at just 5,448 Watts.

It doesn’t take a mathematician to see how skewed these numbers are. Intel, the company with an apparent focus on “the most energy-efficient computing technologies at scale” has missed the mark here with an ancient processing method. Keep in mind though, that Bitmain’s offering is out of reach for the common amateur miner, with a minimum 195 unit order quantity and a required HK3 AntSpace wet-dry stack container clocking in at a $3.4M investment. Further complicating matters for this new unit is Bitmain’s insistence that if miners are used outside of this liquid cooling container ecosystem, they will not honour any warranties or provide service.

Intel’s BMZ1 is only able to produce the hashrate at an increased power consumption of Bitmains S19+ Hydro. Why Intel would even release this product is unbeknown to many in the crypto industry, maybe it is being used as a signal, to show how the times are changing and how Intel are attempting to get on board, but most likely, it is a failed first attempt to do something worthwhile. Power is a big concern in the mining industry, last year the world’s
crypto mining production used more energy than it takes to power Argentina for an entire year. This is why companies are working tirelessly to produce more power-efficient processes. Intel appears to have the right focus in mind. The only hope is that in due time, Intel can prove us all wrong, and actually produce a dramatic change in the power to hash ratio of the industry, and more importantly, availability.

The dark horse in all of this of course is the cost per terahash. If Intel is able to provide an ASIC for around $1,500, severely undercutting the \ cost-per-terahash by the current manufacturers, there is a valid case to be made for Intel’s entry.

Rumours have already surfaced of a BZM2 in production, and Intel has apparently even received orders from large tech giants such as Jack Dorsey’s Block (previously Square). Does this seem like a good investment to make? Depends again on the cost of acquisition. However, increasing interest often drives innovation that thus far, especially in terms of SHA256 Proof-of-Work machines being cornered by less capable marques charging 4-5x MSRP, we welcome Intel to the stage.

Leave a Reply