For me BTC mining with renewable energies might be a chance for the community to restore some decentralisation in the BTC mining area. From what i have read in different threads here, there seems to be some adventurous statements or completely wrong ideas about (PV-) Mining out there. Therefore i'll try to give some guide with a (simplified) calculation attempt.
First of all let's start with some standard mining calculator like
this one, based on hashpower, power consumption, pool fee and energy costs. Obviously those calculator are very simplified and better in discribing the past or the present rather than the future, but despite that we can extract some general knowledge from it.
- Those rewards are extrapolated from the presend, obviously we need some correction here due to difficulty adjustment
- The costs are directly proportional to miner efficiency [J/TH] and electricity costs [$/kWh], no matter which one you optimize, the effekt for present profitability is the same.
- The ratio of rewards to cost is a reliable indicator for remaining lifetime of the miner
Let's have a look on difficulty increase. Regarding the difficulty increase on a log scale you may notice some kind of constant increase. Depending on the considered time period you may have some minor variation.
SourceDifficulty time stamps:
19.06.2015 -> 49,6924 G
19.10.2018 -> 7,1829 T
19.07.2021 -> 13,6726 T
19.06.2025 -> 126,4114 T
Difficulty Devellopment:
120 M: (126411,4 G / 49, 6924 G)^(1/120) = 1,0675 -> +6,75% Difficulty increase every month
80 M: (126,4114 T / 7,1829 T)^(1/80) = 1,0365 -> +3,65% Difficulty increase every month
47 M: (126,4114 T / 13,6726 T)^(1/47) = 1,0485 -> +4,85% Difficulty increase every month
Regarding the rewards its up to you which decrease you feel the most comfortable with, i will go on with 0,96^x, its a little optimistic, but otherwise the following calculations might be even more sobering.
Now lets combine this knowledge with our standard mining calculator results.
For the mining calculator i choose 1TH, 30W, 0% pool fee and $0,03 / kWh. As a result i got $1,50 rewards and -$0.65 costs (current for 30 days). Cost quota is 0,433.
With the Cost quota we can determine the remaining operating lifespan: 0,96^x = 0,433 -> x=log(0,433) / log (0,96) = 20,5 months left.
Additionally we can calculate the earnings value with some
online calculator: Integral (0,96^x - 0,433) for 0..20,5 = 5.
This means that you will rougly earn 5 times the monthly rewards of $1.50 over the life span of this 1 TH miner.
I did the same for $0,02 and $0,01 and $0 per kWh, you can easily do those calcuations for any numbers in excel on your own (remember, it does not matter if you reduce miner efficiency [J/TH] or electricity costs [$/kWh], only the product of both counts).
What conclusions can we take out of this?
- Even if you run your miner 100 years with free energy the expected value of returns will stay below 24,5 x $1,50 = $36,75 per TH
- Lottery Miner with acquisition costs of $50-$100 / TH will never ever reach ROI without extreme Luck, expect to pay 2-3 times your earnings
- Latest Antminer generations with prices at $30 / TH are so expensive to buy that you can hardly manage to reach ROI, you should not buy at those prices as the complete margin has already been skimmed off by the manufacturer
- Instead you should focus on miners with low aquisition costs (e.g. S19 Generation), this is especialle true for mining with renewables, as uptime far below 100% directly decreases your potential earnings
Now lets move on to renewables. I first assume costs of $230 per kWp PV, which seems to be the lower end of possible prices and in most cases only possible due to extensive DIY knowledge. I further assume 25 years lifetime and 912,5 kWh yield per year, wich leads to energy generation costs of $0,01 / kWh. With those energy costs mining should definitively profitable, right? Not really, the weak point is the capacity utilization. With 1kWp PV + 1kW mining capacity the utilization of the miner is only ~10%, much too low for profitable mining. And what if we choose a smaller miner, e.g. 100W? Lets have a closer look.
So 912,5 kWh per year means an daily average of 2,5 kWh per day. When you simplify production and assume a triangular course you get 0 kWh at the turn of the year and 5 kWh per day in midyear. Obviously there are additional fluctuations. On a summer day we may have up to 16h * 0,8kW = 12,8 kWh = ~2,5x 5kWh. We therefore assume a max fluctuation factor (up&down) of 2,5x. But let us stay with the average production triangle first. We cut the triangle into 10 disks of 0,5 kWh which each could be fully consumed by a 100W Miner. In Case of 10x 100W miner, that would mean, that we start with 1 miner and switch on another on energy production is so high that we get into the next disk. While the
first miner has 19% utilization the last one only has 1% and would probably be a heating rod instead of a miner. With smaller miners we can therefore almost double the utilization, but even 19% utilization is to low for profitable Mining.
Lets move on to energy storage. Obviously you can exchange one of those 100W miners with an 0,5 kWh energy storage to consume energy production and provide it to the miner to increase utilization.
Lets consider the following setup: 1x 100 Watt miner + 6 x 0,5 kWh = 3kWh storage + 300W heating rod. The Miner now have a utilization of 19+17+15+13+11+9+7= 91%.
Let us further assume $150 / kWh storage, so you get $230 for PV and $450 for storage. Your invest trippled, so energy costs go up to $0,03 per kWh but with a nice availability.
Back to our profitability chart we can effort following miners:
30 J/TH: 5 * $1,50 / TH * 91% = $6,8 / TH -> e.g. $700 for S19 j Pro 104 TH
20 J/TH: 8,5 * $1,50 / TH * 91% = $11,6 /TH -> e.g. $700 for S19 j Pro 104 TH underclocked to ~60 TH like described
her and
here10 J/TH: 14 * $1,50 / TH * 91% = $19 /TH -> are there any?
The prices for S19j Pro are absolutly realistic, so we are there, profitable PV-Mining is possible right?
But wait, what about
fluctuations? As i described before a fluctuations factor of 2.5x in both directions is plausible.
You can somehow balance out some fluctuations with over- and underclocking. I did a lot of experiments with my S19j Pro (default 104 TH @29,5 J/TH) and reached operation points of +50% Hashrate @35 J/TH, -20% Hashrate @25 J/TH and -40% Hashrate @20J/TH.
You have to keep in mind that high days are less often then low days. For 2 days with 250% production you need 5 days with 40% production to get back to average of 100% production. With underclocking only you may have 2x days 100% Hashpower and 5 Days 60% Hashpower so ~70-75% on average. With overclocking you may have 2x days 150% hashpower and 5x days 60% which results in ~85% avarage but you would need additional invests for better power supply and advanced cooling solutions.
Conclusion:Even with very optimistic Assumptions it might just be enough to reach ROI with a strong tendency to losses instead of gains. There are no buffers for repairs, paid working time or other unplaned downtimes. At the latest when you have a look at losses due to unused opporunities (e.g. trading the energy market or waiting to buy energy storage until price decreases) Mining seems to be a bad idea, as long as you can not produce mining hardware on your own for much less than the B2C prices.
Did i miss something? Would be fun to get a deeptalk here, as i think the topic is very important for decentralisation of miners.
