This article was highlighted for PRO subscribers, Seeking Alpha’s service for professional investors. Find out how you can get the best content on Seeking Alpha here.
Canaan (NASDAQ:CAN) is a Chinese company whose business is selling Bitcoin mining machine hardware. It has no other significant business and doesn’t receive any recurring software revenue. These machines are second rate, behind other bitcoin mining rig manufacturers like Bitmain. CAN’s machines are currently unprofitable to miners based on the current Bitcoin price and electricity rates to run the machines.
To make things worse for CAN and its Bitcoin mining customers, on 5/11/20, the Bitcoin Halving Event occurred. This Halving (or halvening) event literally means that Bitcoin miners will receive half the Bitcoin they normally received from mining. And Bitcoin miners using CAN’s mining machines are currently losing at least $1 per day.
The only other business CAN has, is its artificial intelligence (AI) chip that only made RMB2.7M in revenues ($390K USD) in 2019, or 0.2% of total revenues. CAN’s AI chip has not been able to compete with those of AI chip manufacturer leaders Nvidia (NVDA) and Advanced Micro Devices (AMD).
CAN is on track to be completely out of cash within 1-2 years. Then it will have no choice but to try to raise funds, which will be difficult as it was difficult for it to even get its IPO done. Most professional money managers don’t want to take the risk of investing in a cryptocurrency company due to banking regulations. CAN has 157M shares outstanding, and is currently trading at $5.50 per share. That makes the market cap a whopping $800M+. This is a nonsensical market cap in our opinion. We forecast a 80-90% downside for CAN to less than $1 per share.
Cryptocurrency expert and Bitcoin mining expert Drew Vosk from his VoskCoin channel on Youtube analyzes CAN’s latest AvalonMiner in this video from 5/7/20. CAN sent him their latest miner, the AvalonMiner 1066Pro, for free in order to get a review on the channel. The following are some images of it from the video:
Source: VoskCoin YouTube Video
Vosk shows that the miner is sold by their distributor, BlokForge, for $999 USD. He tries to say nice things about it, like saying it has nice features, professional packaging, and is great from a performance point of view with no stability issues and easy accessibility.
But then he said at time 10:35:
Obviously this miner isn’t super-efficient and it isn’t super-profitable. Unless you have a really really low electric rate, this isn’t going to be a good buy unless your gambling on the future of Bitcoin going up significantly in price.
This video was published on 5/7/20, when Bitcoin was trading at about $9,000. The following is from the video showing the AvalonMiner’s projected profitability over a year that includes the impact of the BTC block reward halving:
Source: VoskCoin YouTube Video
As shown above, the AvalonMiner 1066 mined 0.000886 BTC in one day. The monthly profit is only half the yearly profit, because the Bitcoin halving event hadn’t kicked in yet at the time the video was made. Once it kicks in, it will be a huge negative for mining profits. Why would anyone pay $1K or more for a miner to only make $44 in profit per year? That would take over 20 years just to break even, and that’s only considering electricity costs. It’s not even considering other factors and costs, such as:
VoskCoin published a video on 2/20/20, titled: “What do you need to mine one bitcoin in 2020?” that goes into detail on the difficulty of Bitcoin mining today.
He states in the video that it’s only going to get more and more difficult to mine a Bitcoin as time goes on. The following chart measures how difficult it is to Bitcoin mine:
As shown in the chart above, it consistently has gotten more and more difficult to Bitcoin mine, and will get even harder as time goes on. The website explains that “a high difficulty means that it will take more computing power to mine the same number of blocks”. More computing power requires more electricity.
In the video, Vosk uses the Bitmain Antminer s17e, which was one of the best, most profitable miners today. As shown below in a shot from the video, Vosk mined 0.001028 BTC with the Antminer.
Source: VoskCoin YouTube Video
Recall above, the AvalonMiner 1066Pro only mined 0.00089 Bitcoin, which is about 11% less. And that’s CAN’s latest generation and best miner, while the Antminer s17e is Bitmain's previous generation miner.
The Chinese RMB to US dollar exchange rate is about 7 to 1. The following are CAN’s select financial results over the past three years:
Source: CAN Annual Report
As shown in the 2019 Annual Report, 2019 product revenue (Bitcoin mining machines) of US$200.1M makes up 98% of the company’s revenue. The rest of the revenue is insignificant, with Lease revenue US$3.5M, Service revenue $0.4M, and Other revenue $0.4M.
With the higher Bitcoin price in 2017, CAN was able to sell their machines at a higher price than it sells them for today. Gross margin was 46% in 2017 and 19% in 2018. Bitcoin prices remained low in 2019, thus turning gross margin negative, to -36%.
Because of a sharp decrease in the Bitcoin price, CAN had impairments of inventory and prepayment write downs of US$112M in 2018 and US$104.7M in 2019. This negatively affected the company’s net income for those years. Since the majority of CAN’s US$149M loss in 2019 was from write downs, negative cash flow from operations and investments was about -US$42M for the year.
In the Q419 earnings call on 4/9/20, CAN CEO Shaoke Li stated:
We expect our total net revenues to be no less than RMB60 million in the first quarter of 2020.
That’s only about USD$8.6M guided for Q120. If we assume that level of sales will remain for the entire 2020 year, that’s only about $35M in revenues for the year. That’s a whopping 83% decline from 2019 revenues. Of course, part of the company’s low revenue in Q120 is due to the coronavirus outbreak in China and the rest of the world. Time will tell if the company can improve sales in later quarters with the headwinds of the halving event.
CAN’s cash balance of US$74M at the close of 2019 will likely not last two years before the company is out of cash unless Bitcoin considerably increases in price. Matt D’Souza, co-founder and CEO of crypto mining hardware broker Blockware Solutions, was quoted in this 3/15/20 report that Canaan “may have gone bankrupt” if the company didn’t raise $90 million from the IPO sale, but ultimately, it was “another lemon delivered to investors”.
Over the past few weeks, bitcoin has rallied along with the stock market likely due to speculation regarding the Block (or Bitcoin) Halving Event that happened on 5/11/20. The following chart shows bitcoin’s recent rally over the past month:
As shown in the chart above, Bitcoin's price started rallying in mid-April. That was also about when the stock market started rallying. We believe it also rallied due to the Bitcoin Halving event. Many believe this event increases the value of Bitcoin because miners mine less, therefore less Bitcoin are constantly being added to the circulation, which, in theory, decreases Bitcoin inflation.
This bull case from the halving event is weak in our opinion. And it certainly wouldn’t cause an organic jump in BTC right away. It will likely take months or even years to take effect. The recent jump in BTC was just from speculation and excitement of the halving event, and we expect it to come back down in the near future as the halving event came and went.
CAN has also rallied over the past two months, as shown in the two month chart below:
Source: CAN share price
People think CAN is a Bitcoin play, therefore it is going up with Bitcoin. This is true to a certain extent. However, CAN doesn’t do any of its own mining, like its competitor, Bitmain does. CAN also doesn’t own any Bitcoins. It just makes its revenues by selling the Bitcoin mining hardware. However, its mining machines are currently unprofitable, therefore CAN sells it to miners at below its manufacturing cost. This is why CAN’s 2019 gross margin is negative.
On 5/11/20, the Bitcoin Halving Event happened. Block rewards are cut in half, so miners only get half the bitcoin that they were able to mine before. This is bad for miners, and many will have to quit their business. From CAN’s 2019 annual report:
Additionally, the amount of Bitcoin awarded for solving each block will decline, with the next halving event designed to occur in May 2020.
The Block Halving Event is explained in this Forbes article from 4/27/20.
Bitcoin halving has only happened twice to date. On 11/28/12, the first Bitcoin halving event occured, and the price of Bitcoin increased from $11 to $1,000 about a year later. During July, 2016, the second halving took place and the price of Bitcoin was trading at around $700, and in 2017, the price skyrocketed to $20,000. However, the appreciation didn’t come from the halving, but from wild investor speculation.
This third Bitcoin halving event has seen some lofty predictions on Bitcoin prices. Bitcoin has recently rallied on speculation of this event. However, we believe significant Bitcoin appreciation is unlikely as this event has been talked about for some time in Bitcoin circles. We have spoken with several Bitcoin experts and all have said they don’t believe Bitcoin will appreciate substantially from the upcoming halving event.
With the Bitcoin boom, bubble, and burst in 2017, we doubt it will boom to the degree that it did after the previous halvings. Bitcoin is now already well known worldwide. The “new paradigm” argument that it had in 2017 likely wouldn’t catch on as much again. Bitcoin prices are still mainly held up by speculation, as Bitcoin still isn’t regularly used practically to buy and sell goods and services.
If CAN investors are confident that Bitcoin will rally significantly over the next year, then we believe they would be better off investing in Bitcoin. It doesn’t make sense to invest in CAN stock over Bitcoin right now in our opinion. On that note, if CAN short sellers are concerned about a Bitcoin rally causing CAN to rally, then buying Bitcoin would be a suitable hedge to their CAN short. Even if Bitcoin rallies considerably from today’s level, it’s not a guarantee that CAN will become profitable.
Every model of CAN’s AvalonMiner currently shows a $1-$2 Bitcoin mining loss per day. As shown below:
Source: Asic Miner Value Website
The above is the estimated loss per day of the AvalonMiner, post the Halving Event. This is a real time mining hardware profitability measurement for each mining machine on the market today. These estimates are from 5/12/20 when Bitcoin was priced at $8840. The machine profits change up or down with a change in the price of Bitcoin. Of course, this profitability estimate doesn’t take into account other mining costs that we mentioned earlier in this report, and the initial purchase price of the machine.
The website states that the expenses taken into account are electricity costs. To get the above daily profit, we adjusted the electricity cost to a low $0.08 per kWh, which is roughly the cost of electricity in cheaper parts of China. This is shown on globalpetrolprices.com, China cost per kWh is shown below:
China has cheaper electricity than most countries in the west, so most mining is done there. As stated in CAN’s 2019 annual report, 74.8% of its 2019 revenue was from customers in China.
The default cost of electricity the website gives at the bottom of the page is $0.12 per kWh. That’s about the average cost of electricity in the US. The user can change that cost depending on their location in the world, and the mining machine profitability will change. The default currency is USD, which can also be adjusted at the bottom of the page to other forms of currency.
The Asic Miner Value shows a lot of different Bitcoin mining machines. The only ones that CAN has are the AvalonMiners. The ones that are profitable are primarily those from leading Bitcoin mining machine manufacturer Bitmain.
The following are the most profitable Bitcoin mining machines:
Source: Asic Miner Value
The above profitability estimates were also taken on 5/12/20, when the Bitcoin price was at $8840, and electricity costs are set at $0.08 per kWh. As shown above, most of the profitable rigs are the Bitmain Antminer. Their latest and what will be most profitable rig, the Bitmain Antminer S19 Pro, is scheduled to be released in May 2020. It is featured on their website.
We haven’t found any logical reason to buy Canaan’s miners over Bitmain’s miners. Perhaps in 2017, when bitcoin was hot, Bitmain didn’t have enough miners to sell to everyone. However, now that Bitcoin isn’t as hot, there’s no reason to buy the AvalonMiner.
A Bitcoin mining farm is a warehouse full of mining rigs. Due to the large amount of electricity needed to Bitcoin mine, mining farms are the only way to make mining a slightly profitable endeavor.
Riot Blockchain (RIOT), is a Blockchain technologies company with its own Bitcoin mining operation. It produced approximately 139.28 Bitcoins in Q419. From the latest 10-Q, Riot purchased 4,000 Bitmain Antminer S17 Pro miners from Bitmain in December 2019. On 5/7/20, Riot announced that it has purchased an additional 2040 next generation Bitmain S19 Antminers in April and May 2020. It hasn’t purchased any AvalonMiners from CAN.
Another Bitcoin mining farm company, Marathon Patent Group (MARA), announced on 5/11/20, that it purchased 700 next generation M30S+ ASIC Miners from MicroBT. The company claims these will be much more energy efficient than their previous S-9 Bitmain models.
Marathon is another industrial miner that has purchased one of CAN’s competing miners, and it wasn’t Bitmain. The M30S from MicroBT is more profitable than the AvalonMiner according to asicminervalue.com.
According to the following tweet, CAN has partnered up with Semiconductor Manufacturing International Corporation (OTCQX:SMICY). But SMICY is a lower tier semiconductor company.
It seems unlikely that an Altcoin miner would generate significant revenues for CAN.
CAN has stressed the importance of selling other products besides its Bitcoin mining machines. From the 2019 annual report:
The development of AI technology, especially as it relates to edge computing, and the acceptance of ASICs for AI applications is crucial to our future success in diversifying our product offering.
In September 2018, we released the first generation of our AI chip, Kendryte K210, and we began mass production in the fourth quarter of 2018. K210 is a SoC that integrates machine vision and machine hearing functions. We were the first in the industry to deliver commercial edge computing AI chips based on Risc-V architecture and self-developed neural-network accelerator with outstanding performance.
Already, CAN has shown problems selling its AI chip. The passage above states that CAN began mass production of its AI chip in Q418. However, for the entire 2019 year, CAN only had RMB2.7M (or $400K USD) in sales from its AI chip, which is only 0.2% of 2019 revenues.
CAN hasn’t stated what advantage their AI chip has, or who the end buyer is. CAN’s website with information on their K210 chip doesn’t show any specs.
On another note, the fact that NVDA and AMD don’t sell Bitcoin mining machines tells us that they believe the market isn’t big enough to be worth it to make them.
CAN’s CFO, Quanfu Hong, said in the Q419 earnings call:
Okay. So, for year 2019, the sales for the AI chip is RMB2.7 million, and for 2020 actually, we don’t think the revenue is being most important part. The most important part is that we now already cooperate with many top players in the AI industry.
And our mission has not being changed which is, we hope that the AI revenue in the next three years will payout significant part of our revenue structure. And right now in different areas like the smart control and agriculture and industry, we already collaborated with many major companies. So, we basically are very confident in the future market for AI Chip.
Hong is telling investors that the amount of sales isn’t as important right now, but contacts in the industry and future customers are what’s important. There’s no evidence that CAN has an AI chip that can ever take off and make a significant amount of revenues. Given its AI chip sales currently only make up 0.2% of total revenue tells us that CAN has a likely impossible journey ahead.
CAN confirms this doubt in its 2019 Annual Report:
We cannot predict how or to what extent the demand for our products in the AI market will develop going forward. Furthermore, as ASICs may not develop into mainstream solutions for AI technologies and applications, we might not be able to capitalize on the growth in the market for AI technologies and applications with our ASICs.
However, some investors seem to have confidence in CAN’s AI Chip. According to one person on twitter:
CAN realizes that it must pivot away from Bitcoin mining machines and into other products. It says in the annual report:
We plan to work closely with our partners in product development to enhance our visibility in new market trends and meet customer demand by devoting more resources to research and development. We may also need to recruit more employees for research and development and product development, such as software engineers. We intend to continue to capitalize on market opportunities for introducing new product applications and conduct advance planning for our next-generation products in a timely manner. However, if we fail to penetrate into any of these or other new markets to which we devote our resources, we may not be able to generate returns on our investments and our financial condition could suffer.
However, research and development activities are inherently uncertain, and we might encounter practical difficulties in commercializing our research and development results, which could result in excessive research and development expenses or delays.
Increasing its R&D and hiring more employees, as CAN states it may have to do in the above passage, would increase expenses. In our view, an investment in CAN is the equivalent of investing in an AI chip startup hoping to compete with the big players. Not many tech startups with only $400K in revenue have a $800M market value, unless they have some kind of new innovation. CAN hasn’t indicated any new innovations in its AI chip.
Bitmain, the top Bitcoin mining machine manufacturer, hasn’t made it far with its AI chip. Bitmain announced that they are selling their first AI chip, BM1680, in April, 2018. They only put out a couple follow-up news reports on it in 2018, and we haven’t heard anything about it since then.
CAN has a young executive team. This suggests they lack industry and business experience to know how to compete with the big players in the AI chip industry. All of the executives are in their 30s. From the annual report:
Source: 2019 Annual Report
Compare this to the ages of NVDA and AMD executives, who are in their 50s and 60s.
Source: Yahoo Finance
Advanced Micro Devices Executives
Source: Yahoo Finance
We have not been able to find any reason to buy CAN’s AvalonMiner over Bitmain’s Antminer. The Antminer is shown to be much more profitable even with the halving event.
However, there has been some bullish reports on CAN. This report claims that:
“Its AvalonMiner product line includes a variety of ASIC-based rigs with different levels of computing power, ranging in price from $1,000 to $2,000. All of them are powerful enough to profitably mine Bitcoins after the halving event, making Canaan unique among mining rig sellers.”
We find this statement to not be true. There are plenty of other miners that will be profitable after the halving event, and quite a bit more profitable than the AvalonMiner. Most other miners, such as Bitmain and MicroBT, also use ASIC (application-specific integrated circuit) chips. Therefore, we don’t understand how the AvalonMiner is “unique” as the report states.
Currently, most Bitcoin mining machines, including those that are manufactured by CAN, are unprofitable. With the Bitcoin Halving Event behind us, it is now even worse for Bitcoin miners. Bitcoin currently isn’t appreciating with the momentum that is needed for most mining machines to become profitable. If CAN mining machines continue to be unprofitable, then it will be unable to sell them, and the company will have another big loss in 2020 like it did in 2019. Its other businesses, like its AI chip, are still in the speculative stage with no stated advantage, and currently represent less than 1% of the company’s revenues.
At its current $800M+ market cap, we believe CAN is extremely overvalued. In our opinion, the market cap should be around $100M-$200M. As it’s a recent IPO, the float is only about 30% of the total shares outstanding. CAN has a lockup expiry on Tuesday, 5/19/20. If insiders start selling shares, that would increase the float, and could push the stock price closer to what we believe is its fair value.
This article was written by
Disclosure: I am/we are short CAN. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.