This person who sells mining machines cannot be judged by common sense.

On November 5, 2019 ( EST ), Canaan Inc. updated its prospectus again. Since the "maximum fundraising amount" column in the F-1 form cannot be left blank, Canaan Inc. filled in $400 million. The actual amount raised may be less than this number. For example, 36Kr filled in $100 million, but ultimately raised less than $25 million.

Canaan Creative is the world's second largest supplier of mining machines, adhering to the pure technology route. Bitmain, which ranks first in mining machine shipments, follows the path of "technology, industry, and trade". It not only sells mining machines, but also mines and operates mining pools, and also hoards and speculates on coins. Its ability to "make money" is far better than Canaan Creative.

Bitmain was hurt by the "fork war" and is now caught up in serious internal strife. The title of "No. 1 mining machine stock" will most likely be taken away by Canaan Creative. Canaan Creative adopts an AB equity structure, with founder Zhang Nangeng holding 74% of the voting rights, which can effectively prevent "palace fighting".

A smash hit

Before Canaan Creative released ASIC blockchain computing equipment, miners had gone through three generations of mining: CPU, FPGA, and GPU.

When Bitcoin was first created, there was more meat than wolves, and the computing power required was extremely low. People naturally used their PCs to "mine", and the computing power came from the CPU ( central processing unit ).

"Mining" involves a lot of parallel computing. GPUs ( graphics processing units ) have high throughput and high parallel processing capabilities. The efficiency of "mining" is several orders of magnitude higher than that of CPUs. After the advent of GPU mining machines, CPUs quickly left the stage.

The GPU is only a guest appearance, because it is designed for graphics processing, and many of the built-in hardware are useless for "mining" and are very power-consuming. The active period of FPGA ( field programmable gate array ) is similar to that of GPU. With the emergence of ASIC, both have left the stage one after another.

ASIC ( application-specific integrated circuit ) is a chip specially customized for a specific need, such as dedicated audio processing chip, video processing chip, dedicated AI chip, etc.

The CPU has powerful comprehensive processing capabilities and is an all-round player that can meet a variety of complex needs. ASIC is designed for a specific purpose and cannot perform other tasks.

The CPU is like a Swiss Army Knife, and the ASIC is like a simple bottle opener. The Swiss Army Knife is powerful, but expensive, heavy, and complex. If it is only used to open beer, it is far less efficient than a bottle opener.

As the computing power required for mining becomes higher and higher, electricity becomes the biggest cost. Low-power ASIC mining machines have become the absolute mainstream, and the mines must be located in areas with ultra-low electricity prices.

According to the current circulation price of Bitcoin and its requirements for computing power, even if ASIC mining machines are used, it will not be profitable if the electricity price is higher than 0.2 yuan/kWh. Places with electricity prices that low are all underdeveloped. Wind, solar or hydropower resources are abundant, local electricity consumption is close to zero, and the electricity generated cannot be "connected to the Internet" for various reasons, so using it for mining is also considered "waste utilization."

Canaan Creative was registered in Beijing in April 2013. In September 2012, before the company was established, it released the world's first ASIC blockchain computing device ( using 110nm chips ) "Avalon" and started pre-sale at a price of 9,300 yuan per unit. The so-called pre-sale is actually product crowdfunding, and the terms are very strange, such as R&D may not be successful, success may not be mass production, mass production may not be shipped...

Avalon started shipping in April 2013. As its mining efficiency is hundreds of times higher than that of GPU mining machines, it was once hard to find a machine, and the price was hyped up to more than 200,000 yuan per unit.

The launch of the Avalon mining machine marks the entry of the blockchain industry into the ASIC era, and also means that Chinese companies are at the forefront of the world in independent innovation of blockchain underlying technology.

In the following five years, Canaan Creative developed and mass-produced 28nm and 16nm chips, becoming a top player in the blockchain repetitive computing field.

In August 2018, Canaan Creative released the world's first 7nm ASIC chip and a mining machine based on the chip, Avalon 9, before Bitmain did. The 7nm chip has the highest computing power density in the industry, which means lower costs and lower power consumption.

It is reported that the Avalon A9, which uses a 7nm chip, has a maximum computing power of 20TH/s. At present, most mainstream products on the market use a 16nm process, with a computing power of less than 15TH/s. For example, Bitmain's Antminer S9i has a computing power of 14TH/s.

In 2017, 2018 and the first three quarters of 2019, Canaan Creative has mass-produced more than 150 million ASIC chips.

Why rush to go public?

Going public means sharing profits with investors in exchange for financing. According to the Pecking Order Theory, internal financing ( i.e. cash flow generated by operating activities ) is the preferred financing method, while equity financing is the "most expensive" and "worst", ranking last.

The top players in the mining industry have already made their first fortune, so it is not necessary for them to go public to raise funds and share a considerable portion of their profits with investors. However, the reality is that Bitmain, Canaan Creative, and Ebang International, which are ranked high, are all "active" in the listing process, and they keep trying despite repeated failures.

Take Canaan Creative as an example. It tried to backdoor-list an A-share GEM company in 2016, applied for the New Third Board in August 2017, and applied for the Hong Kong Main Board in May 2018, but all failed. The filing of the prospectus in the United States is the fourth attempt.

The fundamental reason why mining machine manufacturers are keen on going public is that they are clearly aware that the "youth meal" will not last long and they must prepare for transformation early.

First, the price of Bitcoin fluctuates greatly, and the mining machines, mining pools, and trading platforms that are closely related to it are "dependent on the weather" and lack a sense of security. At the end of 2017, the price of Bitcoin was close to $20,000, and fell below $4,000 a year later. The price of mining machines rises and falls with Bitcoin passively, but "when the currency is hot, the mining machines cannot keep up with the increase, and when the currency is declining, the mining machines cannot be sold." As the old saying goes, "A gentleman does not live in the lower stream." Mining machines are the "hard core" of the encrypted digital currency ecosystem, but they are in the most unfavorable "lower stream" position in the value chain.

Secondly, the total amount of Bitcoin is designed to be capped, and the production will be reduced by half every four years, so the mining cost will inevitably increase. If Bitcoin rises to $20,000 again in 2020, the mining cost will be several times higher than in 2017. Mining will eventually become unprofitable, and the business of mining machines and mining pools will be unsustainable. The mainstream of the "post-mining" era is currency speculation, not "mining".

Finally, no matter how you exaggerate, the essence of Bitcoin is a game. One of the prerequisites for online games to last for a long time is the balance between new and old players. If new players only have the "role" of being pressed to the ground and have no power to fight back, they will lose their enthusiasm for recharging. Without the support of new players, the ecosystem will gradually "wither" and old players will lose interest. In the end, the game will be abandoned by both new and old players.

Old Bitcoin players often hold tens of thousands or even hundreds of thousands of Bitcoins, and the cost is even just a few cents. Whether new players work hard to mine or spend tens of thousands of dollars to buy Bitcoins, they are not in balance with the old players. They are basically using real money and legal tender to help the old players. The only hope is that there will be "newer newcomers" to help.

Mining and cryptocurrency trading are ultimately just a game and a dream. The fundamental reason why the mainland and Hong Kong exchanges say no to mining machine manufacturers is that they do not recognize the sustainability of their business.

When that day comes, will the technology accumulated by mining machine manufacturers be worthless? Of course not.

First, in the field of ASIC chips, Chinese companies have the opportunity to "overtake on the curve". Although they are both chips, the design difficulty of ASIC and CPU is not at the same level. Canaan Creative is the first manufacturer in China that can design and mass-produce 7nm ASIC chips. However, it is still a long way from designing and mass-producing 7nm CPUs. Currently, only Huawei in China has truly done this, but Huawei's annual R&D expenses exceed 100 billion ( 101.5 billion in 2018 ).

Secondly, the application prospects of ASIC chips are infinite. In the future, all household appliances, even the "household appliances" mentioned by Uncle Benshan - flashlights, can be equipped with chips, but they must be ASIC chips, not CPUs.

Finally, blockchain technology is the core technology with the greatest potential to trigger a disruptive revolution after steam engines, electricity, information and Internet technology ( McKinsey’s view ). The core mechanism of blockchain is “decentralization”. To realize this concept, the nodes that provide computing power must be “profitable”. Virtual currency, as a reward mechanism for blockchain distributed computing, is an indispensable part.

ASIC chip research and development institutions such as Canaan Creative have achieved the ultimate in high performance and low power consumption, and will have great potential in the fields of blockchain, Internet of Things, and artificial intelligence. It is like top car manufacturers spending money to set up F1 teams, with the goal of achieving the ultimate in engine, transmission, and chassis technology and gradually applying them to mass-produced models.

The ASIC chip field has long been flourishing, and even Gree's "Auntie Dong" is working on chips. Gree must be working on chips specifically for air conditioners, which is theoretically feasible. However, the performance and power consumption are definitely inferior to those of mining machine factories that have participated in the "F1 Formula" competition for many years.

The above are all possibilities. It will take a lot of effort for the mining machine manufacturers to get a piece of the pie of "chips for everything". It is a very important step to seek listing while the mining machine and mining pool business can still make money and open up continuous financing channels.

If the US capital market can accommodate Canaan Creative, then the day when Bitmain and Ebang International come to Nasdaq will not be far away.

It cannot be measured by common sense

In 2017, Canaan Creative's mining machine sales revenue was 1.303 billion yuan. It only sold one model, A7, throughout the year, with 295,000 units shipped at an average price of 4,424 yuan per unit.

In 2018, the sales revenue of mining machines was 2.699 billion yuan. A total of 559,000 units of the three models A7, A8 and A9 were shipped, with an average price of 4,826 yuan per unit.

In the first three quarters of 2019, the sales revenue of mining machines was 945 million yuan. The three models of A8, A9 and A10 shipped a total of 410,000 units ( a year-on-year decrease of 10% ), with an average price of 2,302 yuan per unit, a year-on-year decrease of 56.8%.

The shipment volume and price trend of mining machines are positively correlated with Bitcoin, but mining machines will only catch up two or three quarters after Bitcoin starts to rise. Once Bitcoin falls, the price of mining machines will immediately "dive".

In the first three quarters of 2019, Canaan's A8 series, which uses 16nm technology, still had the highest shipment volume, reaching 266,000 units. The most advanced A10 series has only half the energy consumption of the A8, with an average sales price of 7,556 yuan per unit, while the A8 series is only 1,206 yuan per unit.

Users are reluctant to pay for the latest products ( for example, iPhone 11 sales are only one-fifth of iPhone 10 ), indicating that they still have doubts about the price trend of Bitcoin.

Due to its high price, the A10 series achieved sales revenue of 427 million in the first three quarters of 2019, accounting for 45.9% of total revenue.

The A10 series is expensive but has low shipments, while the average price of the main model A8 series has dropped from 4,842 yuan per unit in 2018 to 1,206 yuan per unit, resulting in a significant drop in the overall average sales price, which was 4,826 yuan per unit in 2018 and 2,303 yuan per unit in the first three quarters of 2019.

In 2017, Canaan's gross profit was 604 million yuan, with a gross profit margin of 46.2%. In 2018, the good times ended, with a gross profit margin of only 18.8%. Interestingly, in the first three quarters of 2019, the price dropped by more than half, and the gross profit margin only dropped by 2.4 percentage points. The drop in the cost of a single machine is amazing. It seems that suppliers are "overcoming difficulties together", which shows Canaan's bargaining power with upstream suppliers.

The three expenses of Canaan Creative are characterized by one low and two highs: marketing expenses are low, which were 14.2 million yuan in the first three quarters of 2019; administrative expenses and R&D expenses were 290 million and 105 million yuan respectively.

The equity incentive cost in the first three quarters of 2019 was as high as 225 million, compared with 14.2 million in the same period of 2018. Excluding the equity incentive cost, the operating loss in the first three quarters of 2019 was 13.8 million.

Frankly speaking, if other companies’ revenue, gross profit, and operating profit were of this scale and trend, they would not be worth paying attention to. However, mining machine companies are an exception. First, Bitcoin may have several more rounds of “big market” and the mining machine market will usher in a “blowout”; second, mining machine companies have accumulated experience in the field of ASIC chips, and there may be opportunities for large-scale monetization.