Mining Breakeven

Another speculative DD on RIOT’s financials. Today I want to try to estimate what is RIOT’s breakeven price for BTC, and thus it’s margin for different BTC price points. Everything here is an estimate, but I try to explain where I come up with the numbers.

Let’s start with RIOT’s published q3 margin:

Increased mining revenue margin, computed as cryptocurrency mining net of cost of revenues of cryptocurrency mining (exclusive of depreciation and amortization), to 76% for the three-month period ended September 30, 2021

...

The average BTC price used to calculate Riot’s third-quarter 2021 mining revenues was approximately $41,837.

If per-coin revenue was $41,837 and margin is 76%, then per-coin breakeven would be 24% or $10,096. Simple.

However, this excludes a lot of real costs; I argue that the $10,096 number is *only* power costs. The rest can be found in the quarterly 10-Q. The other bit of information we need is to sum up RIOT’s published “coins mined” number.

• Q4: 1,355 BTC
• Q3: 1,291 BTC
• Q2: 680 BTC
• Q1: 491 BTC

We can see how power costs have been trending over time by looking at the “cost of revenues” line in the 10-Qs and then dividing each quarter’s value by the number of coins mined:

• Q3: $10,096 / BTC
• Q2: $13,713 / BTC
• Q1: $15,344 / BTC

This at first seems to be trending down nicely, but I don’t think it’s likely to last. Until Q2, RIOT was mining entirely out of the Coinmint facility in NY, which had some additional hosting fees on top of power. Starting in Q2 and extending into Q3, they are expanding, likely exclusively, in Whinstone. At this point, ⅔ or more of their machines are at Whinstone. Their power contract is fixed, so I wouldn’t expect much of a decrease beyond this.

Also as difficulty rises, the hashrate and thus power required per coin will rise with it, so this will probably flatten and start to rise again in Q4 or Q1. More on this below.

The first additional cost is the cost of the miners and other supporting equipment, spread over the usable lifespan of that equipment. RIOT recorded $12.207M in “depreciation and amortization” in Q3 and $5.738 in Q2. Dividing by the coins mined, this is an additional:

• Q3: $9,455 / BTC
• Q2: $8,438 / BTC
• Q1: $5,796 / BTC

At the end of Q3, RIOT reported 2.6 EH and now reports 3.1, so they have deployed 20% more machines, but only mined 5% more BTC. Thus I expect the Q4 equipment cost to come in at 15% higher, or close to $11,000 / BTC.

Lastly let’s look at the big catch-all category “Selling, general and administrative”. This encompasses staff primarily, but also rent and other kinds of “costs of running the business”. In theory these should get cheaper as the company expands. Let’s see how it works out:

• Q3: $40.3M -> $31,221 / BTC
• Q2: $3.5M -> $5,164 / BTC
• Q1: $5.4M -> $11,124 / BTC

It’s unclear what the trend here. It seems like the company is just handing shares out to the CEO and the board. The CEO gets 10 BTC a year too! The big line item is the performance-based incentive plan. From the Q3 10-Q:

During the nine months ended September 30, 2021, the Company awarded 3,925,000 performance-based restricted shares of common stock under the 2019 Equity Plan to employees, which are generally eligible to vest upon the successful completion of specified milestones related to added infrastructure capacity and also adjusted EBITDA targets over a three-year performance period beginning in 2021 and ending on December 31, 2023.

​

So, that’s around $80M in shares at current prices up for grabs over next 2 years (minus what’s already been paid). Or about $10M per quarter, assuming they don’t grant more. Worse even if the stock price actually goes up. I don’t see the per-quarter BTC price of this dropping below $10k and it could easily be much higher.

So we have a Q1 total cost of $32,264 / BTC and a Q3 total cost of $50,772 / BTC!

Even if the Q3 stock vesting was a one-time event, it’s still looking like we end the year at around $10k/BTC power, $11k/BTC machines and $15k+/BTC for employees. That’s $36k/BTC. Given the $41k or so BTC price, it’s only a 14% margin. The cost is also increasing in every category except power. I’m not surprised that power is the only cost the press releases consider when calculating margin.

What about next year?

Let’s talk about mining difficulty. The costs are all tied directly to hashrate. Double hashrate and you’ve doubled your machines and thus doubled your power needs. Maintaining all of those machines requires more people and everything else, so you probably get something close to double the “administrative costs”.

BTC mined is the hashrate divided by difficulty. So if difficulty doubles, all your costs double without your revenue increasing at all.

The above numbers are all operating on Q3 costs since we don't have any Q4 numbers yet, but we know that the Q4 mining difficulty was already about 36% higher than in Q3!

Take a look at this coindesk article: 8 Trends that will shape Bitcoin Mining in 2022. Here's the section titled "Hashrate Doubling":

It’s unanimous; the hashrate for the Bitcoin network will increase significantly next year. Some estimates project it will double

So, by the end of next year, we might expect RIOT’s breakeven cost to be somewhere in the vicinity of $72k / BTC, quite reasonably higher.

It doesn’t really matter how much hashrate they have, if every coin costs more to mine than it’s worth. I know quite a few on here think BTC will be worth $1M some day, and that's fine, but unless BTC price heads up, it'll be interesting to see what happens to RIOT.