Bitmain: IPO and technological challenges

The best two years for the market Bitmain has not released a new chipset, and does not updated its range of equipment for mining Bitcoins that can lead to fatal consequences in such a fast-growing industry. This article tells the story of king Hasrat – from startups to the top of the industry and further to the subsequent gap. Is it possible the revival of Bitmain in the face of growing competition, if the company doing AI and will launch an IPO?

We offer to your attention translation of the article, which was prepared by the Allen Guo of 2048 Capital – an offshore cryptocurrency investment Fund. After the publication of the article to WeChat company Bitmain asked WeChat to remove it. Having been refused, Bitmain began to exert pressure and twice sought the removal of the article. Finally, with some effort, 2048 Capital took control over the account into their own hands and began to share discoveries and insights with the industry. Industry chips for mining is one area where the Chinese chip manufacturers can become world leaders. We hope that these efforts can provide to Chinese companies, miners and investors with new facts and information, as well as to reduce vicious competition in the industry and promote the healthy development of Chinese blockchain market in the long term. Let’s see what Allen is ready to tell us.

In the second half of 2018, the three Chinese companies manufacturing chips for mining, preparing to become a public that is particularly interesting, given the recent China-us duties. Among these companies, special attention is given Bitmain. Being a monopolist in the industry, the company turned to the creation of artificial intelligence (AI) and trying to raise money before launching an initial public offer (IPO).

The purpose of the article

In the past year 2048 Capital is mainly invested in the blockchain infrastructure, public baccani, tools and transactions. To date there has been investment in the miner was not mining bitcoins. Recently, the Fund got interested in IC and mining company Bitmain for the following reasons:

  • I studied digital circuits and systems with low power consumption while studying at Berkeley. After years of working in the Internet industry I suddenly realized that the cornerstone industry of the blockchain is my old specialization – computer chip.
  • Although I was not associated with the chip industry, in recent years many new and old investors sent me the presentation used Bitmain to raise funds, and asked my opinion. I could only tell them their opinion within their own knowledge.
  • With my elementary understanding, I am puzzled by the fact that the flagship product in the mining of Bitcoin was launched back in 2016 For the best two years for the market Bitmain has not released any new chips or miners, and did not update its line of equipment for mining Bitcoins. A rarity for the giants in this fast growing industry.

It turned out that Bitmain preparing for crowdfunding, and in the Internet there are many articles and reviews. But if you take a closer look, you may find that most of these articles incorrectly interpreterpath fundamental logic of the industry circuits. When focusing on only part of a company such as its financial data were rarely analyzed technical strengths, products and expertise. I will write down my thoughts and results of studies over the last few days that, maybe not complete, but as objective as possible.

The core competencies of the company producing chips for mining

Bitmain: IPO и технологические вызовыIn any industry have their own core competencies. In the case of chips for mining there is only one key property, in addition to costs: the rate of energy consumption, i.e. the energy that goes into mining a single bitcoin or the power consumption to maintain the computing power of 1 teraherz (W/TH (watts/teraherz)). Chip manufacturers for any mining company are doing everything possible to reduce the value W/TH, and therefore, the power consumption of its miners.

There are two main approaches to reducing W/TH. The first is to optimize the design of circuits to reduce the power consumption of all components. However, the possible energy optimization is limited. The second approach is to rely on the optimization of the plant’s processes from 28 nm to 16 nm and 16 nm to 10 nm or 7 nm. Optimized process will reduce power consumption but will also increase the costs of chip manufacturing. 28 nm and 16 nm is the width (linear resolution) of the conductors between transistors of the chip. The less linear resolution, the lower the power consumption, but its reduction is limited by the laws of physics. Currently, it is believed that the physical limit is 0.5 nm. TSMC recently launched mass production of chips with a resolution of 7 nm in 2019 plans to begin mass production of chips with a resolution of 5 nm.

The core competence of the chipset manufacturer for mining is to constantly update the design of their chips and improving factory processes to reduce the coefficient of energy consumption. Ceasing to engage in these updates and improvements, the company began to lag behind competitors. Improving the design of circuits depends on the strengths of the team for designing chips, and improving the process depends on the plant.

Analysis of different generations of mining equipment, Bitmain and competitors

1. The miners of the first generation

In August 2013, Bitmain gave in the first production chip for mining Bitcoins, BM1380 using the TSMC process with a resolution of 55 nm. In October 2013, the chip showed a decent test results. Chip was performed using the technological process, two generations of advanced BE100 competitor from FriedCat (烤猫), and achieved lower energy consumption.

Received recognition from the miners, chip has collected several hundred million yuan on sales that allowed Bitmain to accumulate the original capital.

Features of mining equipment of the first generation

Chip Process Efficiency (W/TH)
BE100 130 nm 9353,8
Bitmain S1 (BM1380) 55 nm 2000

Data sources: 1, 2

2. Another major victory

After the initial success of Bitmain fully aware of the advantage of technology and immediately began to design a second chip for mining Bitcoins, BM1382 using the TSMC process with a resolution of 28 nm.

The second chip was put into production in April 2014, and was released in June, showing a slightly better consumption than BE200 from FriedCat (烤猫) with a resolution of 40 nm.

Features of mining equipment of the second generation

Chip Process Efficiency (W/TH)
BE200 40 nm 1125
Bitmain S3 (BM1382) 28 nm 773,8

Data sources: 1, 2, 3

From a technical point of view, in comparison with the BE200 from FriedCat (烤猫), chip Bitmain obvious advantages had. However, as this happens, the team FriedCat (烤猫) chose to focus on providing solutions. With this solution, Bitmain has remained an industry leader chips for mining, and sales of the second chip brought the company hundreds of millions.

3. The mysterious disappearance of competitors on the background of the gloomy prospects for business

Power BM1382 has not met expectations, so Bitmain has continued to optimize the chip using 28 nm process. In October 2014, Bitmain sent into production the third chip, BM1384. Although the consumption BM1384 has been substantially optimized in comparison with the BM1382, it lagged behind the BE300 from FriedCat (烤猫) with a resolution of 28 nm.

Features of mining equipment of the third generation

Chip Process Efficiency (W/TH)
BE300 28 nm 300
Bitmain S5 (BM1384) 28 nm 448

Data sources: 1, 2, 3

In spite of this, Bitmain has ordered a large quantity of boards BM1384 and invested in the production of a large number of miners S5. The production of these miners S5 presumably consumed all previously accumulated profits.

In 2014, the Bitcoin market went into decline. In August 2014, the price of the coins down to about 2000 yuan, but this trend has not ended. In January 2015 the price has fallen below 1000 yuan. Bitmain was no exception and was going through difficult times.

However, this temporary decline marked a significant and dramatic reversal of the state of Bitmain that led to the dominance of the company in the industry. During the last bear market the main competitor Bitmain, FriedCat (烤猫), mysteriously disappeared, and the company ASICMiner (深圳比特泉) was disbanded.

4. Solo performance

In June 2015, Bitmain sent to manufacture the chip BM1385 using the same 28 nm process. In comparison with BM1384 energy consumption and the cost of this chip fell by almost 50%. The main difference between this and all the previous Bitmain chip was a new design procedure, different from the one used by the original design team.

Features of mining equipment of the fourth generation

Chip Process Efficiency (W/TH)
Avalon A6 28 nm 300
Bitmain S7 (BM1385) 28 nm 257,3

Data sources: 1, 2

In comparison with appeared three months later A6 Avalon, BM1385 had obvious advantages in terms of energy consumption and cost. With the growth of the Bitcoin industry in 2015-sale of S7 miner from Bitmain has reached amazing heights. According to the Internet, Bitmain sold 200,000 Asimov that brought 2 billion yuan in revenue with profitability of approximately 80%, and the company’s market share was 80%.

With his miner Bitmain S7 surpassed the entire mining industry.

5. Dominance and prosperity

In December 2015 Bitmain sent to BM1387 manufacture the chip using the TSMC process with a resolution of 16 nm. This chip continues to use the above design procedure, which resulted in the reduction of energy consumption in comparison with BM1385 more than half. It gave Bitmain advantage over the then competitor to the Avalon and the rest of the industry.

Features of mining equipment of the fifth generation

Chip Process Efficiency (W/TH)
Avalon A7 16 nm 150
Bitmain S9 (BM1387) 16 nm 110

Data sources: 1, 2

6. Hidden dangers on the background of the expansion of

The next two years on the market, Bitcoins were observed upward trend, and with Bitmain chip BM1387 continued to dominate in the community of Bitcoin miners.

The above is a price chart of Bitcoin, where it is seen that at the beginning of 2017, the price was 5,000 yuan, and at the peak at the end of that year more than 120,000 yuan.

From the beginning of 2017 first quarter 2018 Bitmain has sold over 3 million miners (including more than 2 million miners S9) and has received sales revenues of about 30 billion yuan with profits of about 60%. Within a few months by Bitmain accounted for more than 10% of the production capacity of TSMC, which makes Bitmain’s second largest client of the company in mainland China after HiSilicon.

However, against the background of continued growth of sales became noticeable two alarm signal.

First: technological stagnation

Over the past two years, the Bitmain products began to lag behind competitors. The power consumption of most of Bitmain products in comparison with competitors more than doubled.

1. Chip for BTC: consumes 40% more energy

Today chip for mining Bitcoins remain a core business Bitmain, but from 2016, all new design of chips for mining Bitcoins was unsuccessful:

  • In February 2017 version optimized for the BTC to 16 nm BM1X89 failed.
  • In December 2017 version for BTC to 12 nm BM1X90 failed.
  • In March 2018 the version for BTC to 10 nm BM1X93 failed.

Thus, Bitmain continues to rely on BM1387 (miner S9) of the sample in the beginning of 2016, while competitors surpassed it in performance. For example, the M10 ShenMa energy consumption can be 40% better than Bitmain.

Chip Process Efficiency (W/TH)
Bitmain S9 (BM1387) 16 nm 110
Innosilicon T2 10 nm 80
ShenMa M10 16 nm 65

2. Chip for LTC: consumes 25% more energy

Chip Efficiency (W/TH)
Bitmain L3 1,6
Innosilicon A6 1,2

Data sources: 1, 2

3. Chip for DASH: consumes 2 times more energy

Chip Efficiency (W/TH)
Bitmain D3 57
Innosilicon A5+ 23

Data sources: 1, 2

4. Chip for ZEC: consumes 2 times more energy

Chip Efficiency (W/TH)
Bitmain Z9 28
Innosilicon A9 12

Data sources: 1, 2

5. For ETH chip: consumes 2 times more energy

Chip Efficiency (W/TH)
Bitmain E3 4
Innosilicon A10 2

Data sources: 1, 2

6. Chip for SC: consumes 5 times more energy

Chip Efficiency (W/TH)
Bitmain A3 1560
Innosilicon S11 320

Data sources: 1, 2

7. Chip for AI: consumes 3 times more energy

Chip Efficiency (W/TH)
Bitmain BM1682 10
DaDianNao 2,8

Data sources: 1, 2

Second: waste of money

If the rumors are true, then Bitmain recklessly squandered 6 billion yuan in an unsuccessful trade.

1. Board BM1385 at 28 nm for a million miners

In the 4th quarter of 2017, with a booming market of Bitcoin, Bitmain sold all my miners. It is reported that Bitmain has suddenly become Board BM1385 in TSMC 28 nm for the production of a million miners new model V9. This step Bitmain shook the entire industry (including TSMC). In the end, BM1385 for more than a year were not issued. It has fulfilled its historical mission and was no longer considered advanced technology. The only reasonable explanation for such huge purchases obsolete chips could be the desire Bitmain to beat the competition, still using the 28 nm process, but had the best design of the product.

Miners soon realized that machine for mining V9 – it’s actually S7, and V9 models are not in demand. In the end, in warehouses collecting dust Bitmain more than a million miners V9 a cost of 2 billion yuan.

2. Risky Board BM1X90 12 nm

It is known that at the end of 2017 Bitmain paid 1 billion yuan for risky untested Board BM1X90 12 nm in the hope to keep up with a thriving market for Bitcoins, expecting a successful trial of issue before sending to mass production. In the end, 1 billion yuan is wasted.

3. Risky Board BM1X93 10 nm for a million miners

Rumor has it that in early 2018 Bitmain has spent 3 billion yuan on a risky Board BM1X93 10 nm for approximately a million miners, which also failed.

The results

Bitmain: IPO и технологические вызовыThe development of blockchain technologies and cryptocurrencies led to a surprising accumulation of wealth over a short period and in a concentrated region. An important cornerstone of confidence that support is the accumulation of technology is a chip to cryptocurrencies, on which depends the future of computing power. Technological innovations are often initially ignored or not recognized as “mainstream”. Bitmain has really contributed to the dominance of the Chinese mining industry. If the international microelectronic giants ignored the cryptocurrency or unwilling to invest in chips for mining, considering them too insignificant, Chinese entrepreneurs are keenly took advantage of the historic opportunity and fully realize the potential of the Chinese global supply chains.

The key to the success of Bitmain that the company used the dividends from production chips, not afraid to invest in advanced manufacturing processes and became a leader of innovation in the design of circuits. The right decisions and opportunities in these two dimensions laid the Foundation of the company’s success over the past four years.

But gradually, the landscape in these two dimensions has undergone significant changes, and today it is completely different. On the one hand, all competitors Bitmain realize the importance of updating the production process, so that advantages in terms of awareness in this area no longer exists. On the other hand, Bitmain has faced stagnation or even regression in their ability to design integrated circuits that became its Achilles heel. What’s worse, the bold investments that Bitmain has brought initial achievements, soon turned into a blind, gambling and even suicide suppression of competitors that harms the cash flow of the company and competition and cohesion of the ecosystem of this young industry.

The core of microelectronic company will always be its technology IC design, and this technology is often developed by a team of founders. Advantages BM1385 and BM1387 from Bitmain was due to proprietary design procedure that optimizes the energy consumption. But, to our knowledge, this technique has not been developed by a team of founders Bitmain. Not realizing the importance of personnel and technology, Bitmain after the departure of the development team ceded the rights to irreplaceable for designing a competitive chip technology.

For companies producing chips for mining, movement toward new areas of AI an inevitable trend, but in this area there are also similar requirements for the technical team. Low power consumption still plays a key role, otherwise, customers are prompted to buy not effective computing power, and the furnace. In the field of AI before Bitmain are the same technical challenges.

I hope that after the successful fundraising Bitmain will return to technological innovations and become the Builder of the ecosystem of the industry and the driving force behind the new digital economy.


Share link:

  • More