Thanks to Elon Musk for bringing the conversation to the forefront of peoples minds again. There have been many discussions about Bitcoin mining over the years, along with Proof of Work mining, and how Proof of Stake is supposedly “better for the environment”.
What we do find is that Bitcoin miners predominantly use renewable energy sources, not just by some small margin but by over 75%! Coupled with others suggesting Bitcoin / PoW mining is pushing forward innovation and growth in the renewable sector, and it seems the assumption that Elon has put forward may be slightly misrepresented.
In fact as someone who’s previously run a small GPU farm of 120-odd GTX1060’s back in 2017 / 2018, the entire time I ran it, the energy utilized came from renewable sources thanks to local provider Meridian Energy.
However, not everyone has the luxury of simply “going green” with their provider of choice, and sometimes coal-burning power is sufficiently cheaper in their area, that a miner will choose it over the environment.
But I want to challenge the assertion that Bitcoin, DigiByte, and in-general “Proof of Work” is a bad thing for a number of reasons.
So let’s start with the obvious question:
How much power does DigiByte consume?
Fair warning: Although I’ve done my best with the math, it’s still not going to be 100% accurate.
It’s actually relatively easy to get a “ballpark” estimate though.
You see, there are a few variables that we can take a well-educated guess at, based on the blockchain data itself.
If we take the difficulty for each of the 5x algorithms (Which we know because that’s how we determine the difficulty for the next block), we can use this to get approx numbers of the hashrate on the blockchain at any given point in time depending on how long it’s taking to find a block for each algorithm (5x algorithms, at 1 minute and 15 seconds average each, gives us a 15 second block timing):
Using this information, coupled with knowledge on each algorithms most efficient miner (hardware):
We can extrapolate and guess the rough power consumption (knowing that not everyone will be using, say, the M30S++ at 110Th/s from 3410w, but some might use the S19 Pro at 110.00 Th/s from 3500w):
Giving us a grand total of 62,407,560 watts. That’s some serious power!
What about Bitcoin though? DigiByte has 40x faster blocks, so you’d assume that all other things being equal (Well, it has just one algorithm vs DigiBytes 5) that it’d be able to process 40x the transactions, making the cost-per-transaction only 2.5%? It’s a little more complicated than that…
Bitcoin, based on a difficulty of 25046487590083, giving us a network hashrate of 163,984.52 Ph/s works out to be around 14.9 million miners, and a power consumption of at least 50,835,199,030w (50.8 TW/h). That is quite a number, approx 814x more than DigiByte!
Now all things considered, given DigiByte at the time of writing has a Market Cap of USD$1.67 billion, compared to the USD$924 billion of Bitcoin, that’s not too shabby, meaning DigiByte has more watts-per-dollar of marketcap securing the blockchain.
So what is the (used) power-per-transaction?
With Bitcoin, given most of the blocks are full at present, it’s relatively straight-forward to work out:
144 blocks per day (86400 / 600), at an average of 1985 transactions per-block, gives us 285840 per-day, comes to approx 4,268 units (kWh) per-transaction.
On the other hand if DigiByte were to be suddenly “at capacity” (as BTC is), we’d be filling 5760 blocks per-day (86400 / 15). This means if we had the same txn’s per-block as Bitcoin (1985), at present, we’d be around 11,433,600 transactions per-day, at 0.13 units (kWh) per-transaction.
That is, again presuming that the hashrate and value of DigiByte didn’t increase of course…
Back in the present world though: DigiByte has done 146,259 txns over the past 10 days to give us an average of 14,625 per-day, or approx 2.5 per-block. Taking 14,625 per-day, at present, gives us 102 units (kWh) per-transaction.
Why the massive difference between BTC vs DGB?
Well quite simply because Bitcoin has a FAR higher value in the block-rewards, in US dollar-value, compared with DigiByte. There is also the fact that DigiByte has a 40X faster block-timing compared with Bitcoin. Both DigiByte and Bitcoin will hopefully implement Schnorr Signatures & Taproot soon, which will improve the on-chain scalability, but estimates are only ~15–20% and there are no guarantees of future implementations naturally.
With Bitcoin, there is 6.25 BTC per-block, plus around 1 BTC in block-rewards every 10 minutes. This comes in at a leisurely USD$50 million a day in block reward subsidy at the time of writing (15th May, 2021 here in New Zealand).
With DigiByte, there is currently 520 DGB reward per-block, and a negligible amount of transaction-fees. This comes to USD$340k per-day in rewards.
As the price of DigiByte / Bitcoin goes up and down, so-too does the hashrate, coupled with gains in efficiency in the mining hardware etc, and this all affects the power consumed on a daily basis, and subsequently the power used per-transaction.
Can’t we just replace an algorithm to be greener?
You see the block-rewards vs USD price of those block-rewards, are what predominantly determines the amount of hashrate a blockchain can sustain.
This is why the likes of Litecoin saw a halving of their hashrate at the same time as there was a halving of the block reward subsidy in 2019, because over the following several weeks, miners realized they could not continue to mine while receiving only half as much in USD-value per-block.
The price of Litecoin also plummeted at the time, and so the hashrate of the network actually fell by 75% by the end of the year. This is why Coinbase increased the number of confirmations required for Litecoin at the end of 2019.
You see, Miners have power-bills to pay, and the cost of hardware to recoup. If mining costs them more than their power-bills, they’ve got a total of four options:
- Hope that the price of the coins they’ve mined increases by the time their power-bill is due (In NZ, that’s the 20th of the following month).
- Hope that someone else shuts off their miners with the difficulty adjustment and that they become profitable. At least, in the case with LTC / BTC and their long difficulty adjustments. This is not applicable to DigiByte where the difficulty is adjusted in realtime thanks to the brainchild of MentalCollatz: MultiShield.
- Find a cheaper source of power, such as moving closer to a power-station, changing providers, though usually anybody who’s mining with more than one or two rigs has taken the time to find the cheapest. As such this isn’t usually “an option”.
- Shut off their miners themselves, decreasing the network hashrate.
Now, as much as I’m a huge proponent of ProgPoW and RandomX both going through our new development process and replacing 2x other algorithms, the fact of the matter is that for the most part it’s not going to change the power consumed per-transaction terribly much, at least not by any largely meaningful factor. ProgPoW and RandomX both utilize largely consumer hardware, so the theory is that a “home” miner is less likely to care about as much direct profitability compared to for-profit semi-professional outfits, though they’re more likely to simply swap to another GPU / CPU mined algorithm instead.
That was what I would do as a small-time miner at least with my GPU’s back in 2017 / 2018 when GPU mining on DigiByte was still a thing. I could sustain slight changes in price profitability, but there would come times when it was significantly more profitable to swap for a day or two to another blockchain before swapping back to mining DigiByte. This equilibrium was caused both by the equation of my hashrate vs network hashrate vs price of DigiByte, and also the hashrate I could attain on another algorithm vs the other network hashrate vs price of the other blockchain.
So it doesn’t really matter, there’s basically miners who are determining their profit based on the cost they’re getting their power for, and as that profitability drops based on the same hashrate vs network hashrate vs price in fiat, they’ll potentially shut off their miners, OR, look to add more / turn shut-off miners back on.
What about HDD mining? What about staking? Aren’t these both “greener” options?
These are both interesting ideas!
There has been a huge boost in HDD mining recently thanks to Chia, which utilizes “plots” on a hard-drive as proof of storage capacity. You take a high-speed SSD with a few hundred gigs on it, it makes the plot over the course of approx 12 hours, and then compresses it down to 100GB for you to store on mass-storage (Such as a hard drive or NAS).
The plots are different from other projects which allow users to backup + store their data on other peoples computers, as you’re not holding other users data. Rather you’re holding pre-calculated plots of information, which you’ll have to regenerate after about 5 years. This regeneration is also something to consider, especially if you’ve already burned out your low-TBW SSD.
It’s something I’ve given a whirl myself, seeing as I have about 10TB on a NAS sitting spare right now, but given it’s still relatively new, I’m still unsure about the whole thing. There’s not enough information about if it’s actually “greener” long-term, though it likely is, but given the amount of hard-drives it’s eating up, critics argue it’s basically just a less-efficient staking, with a “proof of something else”.
It’s also relatively “new and unproven” at this point in the game, and seeing as security must remain at the forefront (A blockchain without security is… pointless?), it’s probably better to err on the side of caution. But, perhaps this is one to watch?
Staking on the other hand, most people will know I’m not a huge fan of. In a MultiAlgo situation it could very well have its place eventually replacing one of the 5x algorithms. However at the very least, during this “distribution” phase of DigiByte through the block-reward subsidy, I think that it’s not an ideal method for fairly distributing the DigiByte asset itself.
This is because with staking, you’ve got to have a lot of money to get more money. There’s a threshold you’ve got to meet (for minimum staking), buying $X worth of DigiByte if we were to do staking, in order to get more rewards. It’s great for keeping nodes online, but that’s where the cost goes to instead of to electricity for PoW: To a VPS provider.
So sure, there’s no large computation going on with staking, but it’s not without compromises to the security / distribution itself that I don’t believe are worth it (at least at this stage of DigiBytes emissions curve). That may change in the future though!
The fact of the matter is that at the moment theres around 12% of the circulating supply sitting on Bittrex (1.8 billion DigiByte), and even Poloniex still have a few hundred million DGB. Both could stake that for themselves if DigiByte were to implement PoS.
It’s not the get-rich-by-doing-nothing scheme many portray it to be.
As such even IF DigiByte were to do staking, there’s a very high chance you’d receive very little unless you had 7-figures worth of DigiByte, and it may not be enough to cover the cost of renting a VPS to run the DigiByte node on. Only the rich would get richer and benefit from staking, whereas with Proof of Work anyone can contribute (especially if we have a CPU / GPU mining algorithm, the barrier-to-entry is almost nil, which increases both security while broadening the distribution of the DigiByte asset itself).
Implementing staking would likely be compromising on security and the decentralization of DigiByte all in one, it just doesn’t make sense, even if it consumes less electricity.
That’s why the energy consumed is a good thing!
It’s this consumption of energy that provides the network with security.
If I have a CPU, and I wanted to try and attack the network on, say, SHA256, I’m going to be horribly out-gunned.
This is because far more efficient hardware exists than I can get my hands on. This is why the block-reward subsidy exists, to ensure that miners have a reason to keep mining when the blocks have very little going on, especially as-is the case with DigiByte right now as the on-chain transaction volume ramps up.
We’re going to need more transactions occurring on the blockchain as we head into the future and the block-rewards decrease. Miners are going to consume power to continously secure the network. This is OK though, because this is how the network protects against malicious actors who would attempt to double-spend.
It’s also this power that prevents any central government from being able to simply “issue more” of the asset, as they can with dollars and cents. It’s what prevents any one company, even a wealthy company such as Apple or Tesla, from being able to control the network and make changes.
And all this power, is 75%+ renewable already, and pushing forward a cleaner / greener future with innovation in the electricity space.
If you ask me, the Proof of Work employed by Bitcoin / DigiByte, is not only the best way to distribute the blockchains respective assets, and at securing the blockchain, but also pushes forward other industries for a brighter future.
So yes, it consumes (predominantly renewable) energy, but it still sounds like a winner to me!