Can the entry of Bitcoin "nouveau riche" into the market support a valuation of 20 billion?

The day before yesterday, another mainland technology company submitted its prospectus to the Hong Kong Stock Exchange. The company is called " Canaan Creative (hereinafter referred to as 'Canaan')". Compared with this difficult-to-pronounce name, it is better known to ordinary people as the manufacturer of the world's second largest Bitcoin mining machine "Avalon".

How profitable is the mysterious Bitcoin industry? The prospectus submitted by Canaan this time is a good opportunity to understand the Bitcoin mining industry.

Huge profits, huge profits

It is said that Bitcoin and virtual currencies are "profitable industries", but how profitable are they? The answer lies in Canaan's revenue, gross profit, and net profit data.





In the past three years, Canaan's revenue growth has been "explosive", with revenue soaring from 48 million yuan in 2015 to 1.308 billion yuan in 2017, with a compound growth rate of 2740%. But this is not the most exaggerated, because the revenue figure is the smallest among the three, and the net profit in 2016 increased by 3091% compared with 2015.



This growth rate is astonishing even in the semiconductor industry. Take Nvidia, whose stock price has risen rapidly in recent years, as an example. Its revenue growth in the last two fiscal years (fiscal year comparison) was only 40%, and its best annual net profit growth rate was only 171%.



What is even more frightening is the proportion of gross profit and net profit in the overall revenue. In 2017, Canaan's gross profit accounted for 64% and net profit reached 46%. To put it in a more vivid way, for every 10 yuan of goods sold by Canaan, nearly half of it can be kept in its own pocket.

Huge profits, really huge profits .



In fact, in the past three years of Canaan's development, expenses have also experienced a rapid growth process. Among the three major expenses, administrative expenses have grown the fastest. This part includes most of Canaan's salary expenses, professional service fees, depreciation, rent, etc., which have increased more than 20 times in two years. However, due to the strong performance of overall revenue and net profit, these are nothing.



The strong revenue also brought Canaan ample cash flow. As of the end of March 2018, Canaan's current assets on the books had reached 2.1 billion yuan. Although the current load was as high as 1.3 billion yuan during the same period, more than 700 million yuan of the funds were actually internal loans.

Judging from the financial figures alone, Canaan will most likely be classified as a "can't miss" stock, but in addition to the financial figures, you should also understand this industry better.

Bitcoin Mining Machines: The Charm of Semiconductors



AvalonMiner A841 Bitcoin Miner

Why is Canaan so profitable? Because the Bitcoin mining industry is extremely profitable. Why is the Bitcoin mining industry so profitable? Because it actually reflects the value of semiconductor technology and the layout advantages of the existing global semiconductor industry.

As a hardware device for "mining" virtual currency, various mining machines (including those other than Bitcoin) do similar things - they prove their "workload" by running rigorous mathematical calculations, and thereby "share" the preset virtual currency with other miners in the entire virtual currency network.

This model was created by the still mysterious founder of Bitcoin, Satoshi Nakamoto. Judging from the nearly nine-year development history of Bitcoin, it is indeed a simple and effective virtual currency distribution mechanism.

Two opposing indicators of chips: versatility and efficiency

Since "mining calculation" has been fixed since the creation of virtual currency (at least for Bitcoin), the trend of "specialization" has become more and more obvious in the history of mining equipment. Specifically, the mainstream hardware equipment for Bitcoin mining has gone through four stages: CPU, GPU, FPGA, and ASIC. The overall trend is that the efficiency is higher as time goes by. At present, the only new products of Bitcoin mining machine companies around the world are ASIC.

The production of ASIC chips is also quite interesting. Mining machine companies are "unique" in the entire semiconductor industry. As typical "Fabless" semiconductor companies, they are mainly responsible for the design of mining ASIC (Application Specific Integrated Circuit) chips. The manufacturing of chips is handed over to wafer fabs such as TSMC, Samsung, and GlobalFoundries. After the wafer fabs produce these chips, mining machine companies will also purchase other electronic components and assemble them together with the chips into the final mining machine.

From another perspective, a large part of the success of mining machine companies is also due to the close cooperation between them and the wafer fabs. If we ignore the purpose, mining chips are even more advanced than domestic mobile phone chips in chip manufacturing process.

Bitcoin Mining Industry Roadmap

However, unlike other well-known "fabless" semiconductor chip manufacturers, such as Nvidia, ARM, and AMD, Bitcoin mining machines are products that "don't have to worry about sales." ASIC chips are used for specific purposes and users. For a group of Bitcoin miners, the parameters of a mining machine, combined with parameters such as Bitcoin prices and electricity costs, can directly calculate the theoretical payback time.

Thanks to the boom in the virtual currency market around 2017, not only was Bitcoin trading hot, but even mining machines were being speculated on. For a period of time, some Huaqiangbei merchants even doubled the price when selling mining machines.

Another point that is different from other semiconductor companies is that mining machine companies, because of their specific applications, do not need to cover a series of troublesome issues such as "Wintel" ecology, software compatibility, hardware iteration compatibility, etc. Even the control of the mining machines is controlled through the Internet.

It is rumored that Bitmain will be a major customer of TSMC this year

This also makes the tasks facing mining machine companies extremely simple, with only two tasks:

• Design efficient mining chips as much as possible;

• Try to find advanced processes to manufacture these chips whenever possible.

Ultimately, this has turned mining machines into a highly profitable and brutal industry. The most important thing is when to "start". It is not difficult to design chips, but the more advanced the chip manufacturing technology, the greater the demand for start-up capital. At the same time, only when the business itself grows to a certain level, will wafer foundries such as TSMC give priority to allowing you to use their new manufacturing process (otherwise there will be a long queue).

This is like a group of people "rolling snowballs" down a hillside. The largest "snowball" at a certain moment is most likely the first one to start. When the gap between the two "snowballs" is large enough, the small "snowball" will be naturally eliminated by the market, and only the front row will survive. And this elimination rate is much higher than that of "traditional" semiconductors.

There is a sense of "winner takes all" in the Internet industry.

The market structure of Bitcoin mining machines

China has long been ranked first in the field of virtual currency mining machines. Canaan's own prospectus quoted two survey results from the famous consulting firm Frost & Sullivan:

• Among the fabless ASIC chip design companies for Bitcoin mining machines, Canaan ranks second in the world;

• The combined computing power of all the Bitcoin mining machines Canaan has currently sold accounts for approximately 19.5% of the entire mining market.

As mentioned in the previous paragraph, Bitcoin mining machines are “simple”, so it is easy for the “winner takes all” situation to occur in the development. The combination of the second place and the 19.5% ratio is the best proof. Because Canaan is also a competitor of all other mining machine companies in the world, Bitmain has a very high market share.

Since the data cited in Canaan's prospectus is not public, we are unable to know Bitmain's specific market share. According to an estimate of 80.5% of the market share outside of Canaan, Bitmain's size is likely to be more than three times that of Canaan (60%).

Global Bitcoin mining pool share

As the enemy of Canaan and other mining machine manufacturers, Bitmain not only produces mining machines, but also owns three mining pool platforms: BTC.com, AntPool and ConnectBTC. The total computing speed of these three mining pools alone has reached 42% of the computing speed of the entire Bitcoin network, which does not necessarily include its own free mining farms established in various parts of China and overseas.

In terms of products, Canaan and Bitmain are also each other's strongest rivals. Comparing the latest generation of mainstream products of both parties, Antminer S9i (Bitmain) and AvalonMiner A841 (Canaan Creative), the Bitcoin mining speed and power consumption of both parties are very similar. However, in terms of price, the third-party price of Avalon is 8,300 yuan, while that of AntMiner is 6,500 yuan, but the former is in stock and the latter is out of stock.

The addition of professional mining machines has greatly increased the "difficulty value" of Bitcoin mining, and some professional mining machine manufacturers have been eliminated in the competition.

Another trend worth mentioning is that in addition to Bitcoin, there are many other virtual currencies that can be obtained through mining. Canaan also revealed in its prospectus that it will start producing ASIC chips for another cryptocurrency in the fourth quarter of this year.

This is why Canaan also revealed its next phase plan in its prospectus: the edge computing artificial intelligence chip KPU, which aims at voice and image recognition and IoT applications in smart homes, smart cities, smart monitoring and smart toys. The time plan is also the fourth quarter of 2018.

This positioning is actually closer to Cambrian, while Canaan’s main rival Bitmain has chosen to focus on the cloud. It is difficult to predict how much Canaan can achieve in this direction at present.

Of course, there are doubts, such as Canaan's technical reserve capabilities. According to the industrial and commercial registration information, Hangzhou Canaan Creative currently owns 16 patents, and the related Beijing Canaan Genesis owns another 9 patents. However, these 25 patents are all centered around Canaan's Bitcoin mining machine products, and are mainly design patents, not general patents.

According to the data in Canaan's prospectus, the company currently has a total of 94 R&D team members and 68 engineers, which is a very insignificant number among traditional semiconductor companies. According to rumors, the R&D team of Canaan's main competitor is several times larger than that of Canaan.

Valuation Conundrum

To be fair, Canaan's success is entirely dependent on the popularity of the entire virtual currency market and the power of semiconductors themselves. But at the same time, the premise of these huge profits is also subject to "sudden changes". Canaan specifically emphasized a risk in its prospectus: "Most of the revenue comes from domestic sales, and any lack of supervision may have a negative impact."

In addition to regulatory impacts, Canaan's "single product line under a single application" is also a big problem. Its prospectus specifically states: "In the foreseeable future, it will still rely on the sale of Avalon Miner to generate the vast majority of its revenue."

Judging from the development laws of the "traditional" semiconductor industry, whether a semiconductor company can develop well in the market for a long time does not depend on which product it produces, but on things such as patents, intellectual property rights, and software and hardware ecosystems. Mining machine companies are obviously not among them.

This also brings up a big problem - it is extremely difficult to value Canaan. Let's not value it first, but see how it calculates itself.

The products of general semiconductor companies are usually consumer goods, commercial solutions, infrastructure, etc., while mining machines are more like a physical "investment product" with high risk and high return volatility, but Canaan's "investment products" are created through semiconductor technology.

According to the share transfer records in the prospectus, Canaan made a capital injection in April 2017 and issued new shares at the same time. Ningbo Zhuoxian purchased 4.6712% of the shares at a cost of 150 million yuan. Based on this, Canaan's valuation was about 3.211 billion yuan at the time. With reference to Canaan's profit of 52.54 million in 2016, the overall P/E ratio is about 61. As Canaan's revenue and profit increased in 2017, this ratio may drop slightly, but it is certain to exceed 50.

The P/E figure had been debated before Xiaomi went public. Canaan's P/E figure is over 50, which is comparable to most Internet companies. Its revenue and net profit growth are indeed far higher than those of Internet companies, but its risks are not comparable to those of ordinary companies.

Working backwards, Canaan's actual profit in 2017 was 361 million yuan. If it were still calculated based on the previous P/E ratio, Canaan's overall valuation would be close to 20 billion yuan.

On one hand, there are huge profits, on the other hand, there are risks, and semiconductor technology is behind it. How will the market view Canaan's 20 billion yuan valuation? Is it willing to hold on to this Bitcoin upstart? Let the market give us the answer.