Very original thought...
I think in the near future (10-50 years) ASICs at 10nm-1nm will be energy efficient enough.
Also, I do not think that Bitcoin requires infinitely increasing number of nodes to function, only needs large number of nodes with sufficient hashing power to prevent someone from mounting a 51% attack.
The currently established model, having few large pools containing majority of the hashing power, I think will be an adequate match to the network's growth challenge to quite a degree.
I don't see how quantum mechanical effects could be suppressed to make 1nm technology viable. Expect it to stop on a few nanometers.
Bitcoin will need to keep growing its computational power because the bigger it gets the more value it has to succeed in a 51% attack. So the pressure to break the network will increase with its market cap.
It is effectively an arms race.
So while it is true that we only need enough power to prevent an attack, the attack severity will also increase.
But i was interested in the
energetic picture.
Not sure how the model of pool operators helps clearing that up.
I don't see mining as a real issue. Probably with very wide adoption one party can't easily reach 51% and most of mining is done, by financial institutions to keep their fees down.
Real issue in my mind is the cost on network from traffic viewpoint. If individuals can use their own clients and everything is highly connected the needed throughput could be quite large. Though likely even this isn't an issue. 10MB blocks is around 1MB/s which is 1,3Mb/s. For 70 000 000 clients(1 client for every 100 people) this is 91.4Tb/s. Which is quite large amount of traffic, on other hand it could be done by using such techniques as multicast... So not bad for future...