I’ve found Jon Koomey’s work helpful in this domain too.
Implausible projections overestimate near-term Bitcoin CO2 emissions
https://www.nature.com/articles/s41558-019-0535-4
Common mistakes that people not familiar with the subject matter make (i.e. common assumptions about the technology, the issues with projecting forwards by 5-10 years in very uncertain fields, etc.)
Does not compute: Avoiding pitfalls assessing the Internet’s energy and carbon impacts
https://www.sciencedirect.com/science/article/abs/pii/S254243512102117
The web3 life itself site covers this too, although it doesn’t really dive into the energy market dynamics:
Finally, although the author obviously has his opinions and priors, this post from Luke Plant is accessible and detailed about the implications of many of the aspects of the technology.
In particular, one of the key ideas that I haven’t so well articulated elsewhere is how the mechanism for securing cryptocurrencies while distributed relies on consensus across all the nodes.
If I have some Bitcoin, and want to transfer it to you, and we are both sitting in the same room and have all the computer and network hardware we like, it is still impossible for me to transfer ownership to you without an internet connection. We have to connect to the Bitcoin network, add our transaction to the public, shared list of all transactions, and wait for confirmation. So, from the perspective of a user wanting to make a payment, it’s better to think of Bitcoin as a large, distributed but centralised system rather than a decentralised one.
Source: The technological case against Bitcoin and blockchain by Luke Plant
Why the distributed but centralised framing is important when talking about accounting for crypto
I think this last part is key. The distributed but centralised framing is important when thinking about the environmental impact associated with a single coin. There’s an ongoing energy cost to maintain the security long after it’s minted, a bit like how there’s an ongoing cost to store something on say… Amazon S3 object storage.
With the Amazon and storage example, there is single entity that is responsible for keeping the files stored safely and securely, along with a single entity notionally responsible for the emissions associated with storing those files. (it gets a bit more complicated than that with GHG scoped emissions, but generally speaking the 3 scopes should cover all this responsibility for emissions).
However, with a proof of work based cryptocurrency, you’re relying on every single node connected to the network that is working to maintain consensus as the mechanism to secure the value allocated to the coins being minted. If every node is necessary to safeguard the coins value, you can argue that the a portion of the emissions from every single node required to maintain consensus ought to also be attributed to the coins too.
The implication here is that over a coin’s lifetime, there is an ongoing amount of energy required to keep it secured via this consensus, and if you want a coin to be green, you need to be able to account for all the energy use attributable to it since it was minted, and then keep accounting for it each year that it exists going forwards.
This last part is the thing I’m looking for some comparisons to modelling wise, as it’s a bit of leap mentally from how I see people talking about green crypto - most of the time, there’s this assumption that a one-off intervention to account for the energy associated with minting a coin is enough, then you’re done.