Due Diligence of Bitcoin Mining stock valuations: RIOT and MARA
Unlike the 2017/18 rise of Bitcoin, the 2020/21 Bitcoin price action has been largely driven and influenced by institutional demand.
With institutional demand comes institutional products. We have seen bitcoin derivatives, bitcoin trusts, and more recently a new way for exposure to bitcoin price action: publicly traded bitcoin mining companies.
Lets begin with a quick summary on bitcoin mining.
What is Bitcoin Mining?
To mine a new block on the bitcoin blockchain you must find a number called a nonce. The cryptographic combination of the nonce + next block content must be numerically smaller than the network’s difficulty target (more on this below). Generating a new nonce is extremely computationally expensive. There is no shortcut, miners must try calculations over and over again until they brute force the correct result. They are trying to find the answer to a math problem; a nonce that hashes correctly with the SHA-256 (cryptographical hashing function).
The bitcoin network adjusts how many computations must be done on average to find a valid nonce every 2 weeks. This is called the difficulty. The aim is to have a new block mined approximately every 10 minutes. This means that if the mining power is doubled, the difficulty is made 2x harder. Double the energy for the same amount of bitcoin.
We can see an estimate for the total hash rate being applied to the bitcoin network today. Notice that it is about 150m TH/s right now (150 million trillion nonces tried every second). So for every block (every 10 minutes) 9*1022 attempts are done (90 sextillion) until one lucky participant finds a valid nonce and can “mine” the next bitcoin block. In doing so they get rewarded 6.25 newly generated bitcoin + some bitcoin in transaction fees from users of the network.
You can run code on your computer to try and find the next bitcoin block. In fact, this is how all bitcoin mining used to be done. However, these days you will end up spending more on electricity than you mine in bitcoin. This is because people have developed special ASIC chips designed specially for brute forcing the SHA-256 function. These chips are a factor of 100 times more efficient than your computer.
Mining on an ASIC Bitcoin Miner
The biggest producer of ASIC Bitcoin Miners is a Chinese company called Bitmain. They produce miners, run many of the miners, and even run many mining pools where people can collectively search for the next block and split the profits.
Bitcoin mining becomes incredibly profitable when there is a rapid price increase in bitcoin. This is because there is a lag while new bitcoin mining hardware is being built and deployed to capture excess profit.
So lets take a look at how profitable it is to run an ASIC miner today.
The Antminer S19 Pro is one of Bitmain’s latest machines. It can run an impressive 110 TH/s and uses 3250W of power. On their website Bitmain is listing them at $3769 per machine but in bulk companies can buy them for as low as $2333 (notice they are sold out until at least August).
- $2333 machine cost
- Using $0.08/kWh electricity costs (benchmark for electricity cost in China) we get 3.25kW * $0.08 = $0.26 per hour in electricity costs. Assuming we run 24/7 this gives us $0.26 * 24 * 365 = $2278 in electricity costs per year
- And how many bitcoin will we mine? Well at 110 Th/s on an ASIC machine we are capturing 110/150,000,000 = 0.0000733% of the total bitcoin network per machine at current total hash rate. 52,560 bitcoin blocks mined per year; 6.25 bitcoin reward per block + ~1.25 bitcoin in transaction fees (avg right now) -> 7.5 bitcoin per block * 52,560 blocks -> 394.2k bitcoin per year for the entire network. 0.0000733% of this is 0.289 bitcoin per machine per year = $14,450 revenue (at $50k per bitcoin)
- $14,450 – $2278 = $12.2k profit per year
You can play with the parameters to figure out different profit levels. For example, if the total hash rate doubled to 300m Th/s, profit per machine would drop to $5k per year.
Also note that transaction fees fluctuate over time depending on how many people are actually using bitcoin (sending and receiving transactions). See historic transaction fees.
And how long can you run a machine for?
Running so many calculations slows a machine quickly. Additionally, new faster hardware is introduced making past hardware unprofitable fairly quickly. As a rule of thumb we can expect a machine to be profitable for about 2 years. This is also the rate at which companies mark the depreciation of miners on their balance sheet.
In Year 2 the machines become much less profitable than initially.
There have been studies to see where bitcoin mining is occurring through IP address analysis of miners working in pools. The University of Cambridge has a real-time map in which we can see that in 2020 mining was:
|Country||% Total Hash Rate|
RIOT and MARA
Now that we have a basic overview of how bitcoin mining works let’s look our new investment phenomenon: publicly traded NASDAQ listed bitcoin mining companies.
Incorporated in July 2000, the company was renamed to Riot Blockchain in October 2017. The company now exclusively mines bitcoin and has a partnership with Coinmint LLC in New York who operate the miners. RIOT was a penny stock for most of its existence and was irrelevant until mid 2020.
How many miners do they have?
- Current status of pre-Dec 2020 orders: 11.5k miners in operation and 10k more to be delivered (funded from $100m secondary offering in Oct 2020)
- In December 2020 their stock price rocketed alongside bitcoin ($3->$15 in 1.5 months). They took advantage of this and ran a secondary offering raising $200m. With this money they ordered 15k more miners for $35m. Expected delivery starts in May 2021 with the bulk to be delivered in October 2021.
Summary: 37,640 machines at full deployment. 11.5k machines today and won’t be running the rest until Q3/Q4.
So ignoring the cost of the machines:
- Operational 11.5k machines at $12.2k profit per year per machine = $137m
- Full capacity (Q3/Q4) 37.6k machine at $12.2k profit per year per machine = $459m
- Maybe $150m in cash/btc left over from secondary offering and past assets
Note: They may get cheaper electricity costs through their partnership and claim as low as $0.0014 per kwh in some releases.
Incorporated in 2010, the company has a wild history. It first engaged in the exploration and development of uranium and vanadium, then real estate in Southern California, and most recently patent trolling. The company has some questionable leadership and has also been a penny stock for most of its existence. In 2019 (I think?) the company began delving into bitcoin mining.
In September 2019 they purchased 6k S9 Bitmain miners (13.5 TH/s) for $4m. Notice that at 13.5 TH/s: these machines are far from profitable today (I’m not even sure they were profitable back then!).
So more recently they own:
- Current status of pre-Dec 2020 orders: 6.5k miners in operation and 30k more to be delivered (funded from $100m secondary offering in Aug 2020)
- In December 2020 ran another secondary offering after their stock price went from $2->$14 in a month. They raised $200m. With this money they ordered 70k more miners (delivery of 7k in July, 63k in December) for $168m
- In January 2021 they ran yet another secondary offering raising $300m more. They used $150m of this to buy 4.8k bitcoin at $31k per coin
- Operational 6.5k miners = $79m per yr
- Full capacity (Dec 2021) 103k miners = $1.25bn per yr!
They also made ~$90m on their bitcoin holdings so far ($150m -> ~$240m)
Impressive numbers right? Well..
Too good to be true?
So what are RIOT and MARA trading at you may ask?
With the recent jump of bitcoin to $50k, RIOT and MARA have rocketed.
- RIOT closed at $77.90 per share, a market cap of $5.3bn
- MARA closed at $47.90 per share, a market cap of $4.5bn
So is it because they have such huge potential? If RIOT gets all of its hardware delivered is it really going to be making $459m per year? And MARA $1.25bn per year!?!
Well, probably not..
- As bitcoin price goes up, mining becomes VERY profitable and there is a rush to build and deploy more hardware. From Oct to Dec 2017 the price of bitcoin famously went from $4k to $20k. Total hash rate went up about 60% during this time. However, it went up a massive 600% over the following 9 months as new miners were produced and deployed with a lag.
- As the price of bitcoin capped out and then dipped mining soon became unprofitable at such a high total hash rate and miners lost a lot of money.
- Both RIOT and MARA are pending receipt of the majority of their miners and won’t be taking delivery until late 2021/early 2022. What will the network’s total hash rate be by then? What will bitcoin price be by then? Chinese miners won’t be sitting by idle.
- And we ignored the cost of the machines! In the future RIOT and MARA will need to buy new machines. Regardless of the market the old ASIC miners will quickly trend towards unprofitability. Bitcoin is closed system; There will only be a maximum of 328.5k + tx fees of bitcoin mined per year. In fact, this will reduce in 2024 with the next “halving”. A company cannot reliably mine more bitcoin/capture more of the network because if they are finding it profitable to increase their hash rate then you can be certain every other miner is doing the same. How many $100m secondary offerings can they do for this free lunch?
- Take a look at the past financials of these companies. RIOT mined an “impressive” $2.4m worth of bitcoin in Q3 2020. Their costs were $1.3m in electricity/rent/etc. + $1.2m depreciation ($2.5m!). Once you add in their $2m of salaries and marketing they had an operating loss of -$2.1m for the quarter. Their bitcoin did appreciate by a cool $385k meaning their net loss for the quarter ended at -$1.7m.
- Of course their next quarter should look wildly better thanks to bitcoins Q4 2020 price action. But this is not a business that is massively profitable under normal market conditions.
The bull case:
- Bitcoin keeps rocketing and never comes back. The total hash rate doesn’t catch up and $50k btc becomes $1m btc. MARA’s $1.25bn revenue next year becomes $12.5bn before ASIC miner production catches up. Investors doubled their money.
Of course, if it becomes so profitable one wonders if Bitmain decides it is better value to use the miner’s themselves and breach their contract.
Simply stated, when you invest in a RIOT or MARA you are investing in the profitability/unprofitability of bitcoin mining over the next 1-2 years.
It could be hugely profitable, or it could crash and burn like it did in 2018 and you see no returns.
In this instance, you are paying the executive team of MARA and RIOT to buy the hardware, plug it in, and sell any bitcoin for you. There isn’t much more to it than that – you could buy the hardware yourself and run a few machines if you found a good electric company to partner with for cheap electricity.
So perhaps a good way to see what kind of premium you are paying to own bitcoin miner exposure through these NASDAQ companies is the Q Ratio. The Q ratio is the market cap of the company divided by the replacement costs of its assets.
- Market Cap $5.3bn
- Current btc miners 11.5k ($27m)
- Ordered btc miners 26k ($60m)
- $150m cash remaining (guess)
- A whopping Q ratio of 22.4! And 63% of that is cash, 25% in orders!
- Market Cap $4.5bn
- Current btc miners 6.5k ($15m)
- Ordered btc miners 96.5k ($225m)
- $240m in bitcoin
- $100m cash remaining (guess)
- A “value” Q ratio of 7.75!
With this approach, RIOT is trading at a 22.4x premium today and MARA a 7.75x.
A criticism of this valuation methodology may be that it is justified paying a premium when orders for new hardware are so backlogged. Just remember that these companies are also stuck in that backlog and won’t have most of their machines online until 2022. Is it really worth such a premium?
Other factors to consider
These companies have questionable management. The CEO of MARA has awarded himself compensation in excess of $300m over the last few months. Realise that the market cap of MARA was less than $300m just 3 months ago.
Both companies are quick to do secondary offering after secondary offering at such high prices. MARA has been more aggressive recently and the downward pressure of insider selling and secondary share offerings has resulted in the price moving from $20->$48 in the time RIOT has gone from $20->$77.
A read through both of their 10-Qs will give you paragraphs of legal claims, shareholder class actions, and past investment write-offs due to corruption/shady circumstances.
These companies have no moat. They have no IP. They have no key talent hires. They are simply ordering ASIC miners, plugging them in, and having the biggest pay-days of their lives.
So is there a trading strategy I would recommend?
The challenge with being a bear in this market is you have to seriously consider whether the risk/reward is there for betting on the downside.
Right now the market doesn’t care for your fundamentals. Bitcoin has passed $50k, the narrative is that its going to $100k. The market expects these stocks to continue to trade at multiples and be 10x higher in a month.
It sounds ridiculous but it keeps happening. It is far too dangerous to be short these stocks. Or worse, short calls and you could be bankrupt on 1 intra-day move.
So that leaves you with puts. Average implied vol nearing 300% does not make them cheap.
If you truly have conviction that there will be a correction in the near term then consider weekly bear spreads.
Longer DTE plays are insanely expensive and run the risk of the many secondary offerings eventually justifying a price higher than your price target.
Another interesting idea would be a relative trade. If you could find exposure to bitcoin mining at <20x premium your long run earnings could net out. However, all publicly traded companies are similarly ludicrous so you’d have to buy a stake in btc mining privately or try and own some ASIC miners yourself.
More simply you could go long bitcoin and short RIOT/MARA. This should converge in the long run. The risk is that short term the short position will be impossible to risk manage and there is always the risk of more secondary offerings.
Buy as far OTM weekly calls as you can at 400% IV and hope that momentum market continues. Fundamentals are for the 2010s.
Both RIOT and MARA’s stock price spent some time consolidating down from ~28 to the 16-22 range when bitcoin corrected back from its initial $45k run up back to to $30k.
It is possible they will again if bitcoin has a pullback. However, the price action of these stocks has diverged even from bitcoin for the time being.
Ultimately the price action we are seeing now is purely momentum/hype/technical and probably retail driven. Long term and as far as future earnings go this play is a loser. But be careful; that doesn’t seem to matter right now.
Disclaimer: This is not financial advice; do your own DD and trade at your own risk. Positions: I hold 1 60/50 put debit spread for Feb 26th for each company.
TLDR; Bitcoin mining stock valuations are about as crazy as the valuations of all growth stocks in this market. Fuelled by retail demand of people who don’t know what a blockchain is but their crypto-millionaire friend told them its revolutionary. “The greater fool” investing 101.
Disclaimer: This information is only for educational purposes. Do not make any investment decisions based on the information in this article. Do you own due diligence or consult your financial professional before making any investment decision.