Can Bitmain Go Far with Threats from His Rivals (I) – Friedcat and Avalon
In the second half of 2018, the China-based miner manufacturers – Bitmain, Canaan and Ebang are gearing up to go public. Apart from the big three above, latecomers in this space like Shenma, Bitfily and Innosilicon are sprinting up rapidly, someday, maybe they could elbow into the new top three.
As the world’s undisputed leader in cryptocurrency mining hardware manufacturing, Bitmain has recently concluded its $1 billion pre-IPO financing led by Tencent, Softbank and CICC, which boosted its valuation up to $15 billion.
While in the best of the past two years, Bitmain, as a fast-growing company, failed to roll out successful products after Antminer S9. It triggers our exploration on the state quo and future of Bitmain, as well as the entire crypto mining industry.
As space is limited, we’ll elaborate on it in two parts. This is the first part about the rapid rise of Bitmain, and the second article will go details about limits and potential risks faced by Bitmain.
Core competency of a miner manufacturer
Power consumption is an essential indicator apart from hashrate for a crypto miner, that is, the electric power needed to mine a bitcoin or watt per trillion hashes (W/Ths). Mining hardware makers are trying everything they can to lower it, as the savings in electricity bills could be a direct source of profits for miners and would, in turn, boost sales for them.
To lower the W/Ths, one may either optimize the chip design, of which the effect is actually limited, or count on chip foundry’s process improvement like the latest 7nm process which may weed out earlier 16nm or 28nm. Nevertheless, the second choice means the cost of chips will also rise sharply along with the hashrate increase.
*In this case, 28nm or 16nm refers to the width (wire width) of the wire connecting the transistors on a chip. The smaller the wire width is, the less power it consumes, and the better it could perform. While the reduction of the wire width is ultimately limited by the laws of physics. Currently, 0.5nm is considered as a physical limit. TSMC has just begun its 7nm mass production, with 0.5nm scheduled in 2019.
In this sense, the core competency of a miner manufacturer thus lies in the cost cuts and power consumption reduction. The optimization of chip design relies on the R&D team, and the improvement of process counts on the chip foundry. Bitmain’s co-founder, Micree Zhan is a chip designer and that explains why the company’s products made an instant hit in the market in early days.
An overview of Bitmain’s miners
- Hit the market – S1 with 2,000W/Ths
Bitmain’s first bitcoin mining chip BM1380 with TSMC’s 55nm process at 2,000 W/Ths was for tape-out in August 2013. Compared with its back-then rival, Friedcat/Bitfountain BE100, its chip process is two-generation ahead with much better performance in power consumption. Favored by miners, BM1380 brought Jihan Wu and Micree Zhan their first hundreds of millions.
- Keep the buzz – S3 with 774W/Ths
After S1 became an instant hit in the market, seeing the great profits advanced process could bring, Bitmain then rolled out its second bitcoin mining chip BM1382 with TSMC’s 28nm process to keep the buzz going in April 2014. Compared to BE200 equipped with 40nm chips, Bitmain still has an advantage in power consumption, though the gap between them was narrowing.
Technically, the advantages of Bitmain’s 1382 chip over BE200 were not that obvious. However, the Friedcat team made the wrong decision to have others to make mining machines, allowing Bitmain to continue its buzz and its sales continue to spike.
- Look bleak – S5 with 448W/Ths, lagging behind Friedcat
As its second chip failed to achieve the W/Ths as expected, Bitmain continued the optimization with the 28nm process. Six months later, its third mining chip BM1384 were released in October 2014. The power consumption of S5 equipped with BM1384 has been greatly improved compared to S3, but it lagged behind Friedcat’s 28nm BE300.
Even in such a circumstance, Bitmain still placed an order to purchase a large number of BM1384 wafer and invested in the mass production of S5 miner. According to sources, the production of the S5 miners has consumed almost all of the profits accrued from previous products.
The whole year of 2014 witnessed bitcoin going downward. By August 2014, it had dropped to 2000 CNY/BTC (roughly $300) and continued the fall; by January 2015, it had dropped to a record low of 900 CNY/BTC (roughly $150). The future for Bitmain looked bleak amid the anxiety of the entire industry.
However, this bottom has dramatically become a historical turning point for Bitmain in dominating the crypto mining industry. As at this slump, its biggest competitor, Friedcat, mysteriously disappeared, and his company, Bitfountain, was disbanded.
- Dominate the market – S7 with 257W/Ths
In June 2015, Bitmain developed 1385 chip based on the 28nm process, cutting the cost and power consumption by nearly 50%. Sources had it that the biggest difference between the chip and its predecessors was a new design methodology, which in fact did not come from Bitmain’s original chip design team.
Compared to Avalon A6 which came out three months later than Bitmain S7, S7 outperforms Avalon A6 both in costs and power consumption. Together with the bitcoin market going up, Bitmain S7 sold quite well – 200,000 units, total sales 2 billion CNY, 80% profit ratio and 80% of market share.
- Compete with Avalon – S9 with 110W/Ths
With TSMC’s more advanced process 16nm, Bitmain rolled out BM1387 chip for S9 which achieved over 50% reduction in power consumption, further distancing itself from its second rival Avalon and taking the lion share in the market.