> First, when the network space is growing exponentially, the value of preexisting work is decaying exponentially. I do grant you that completed work is nonzero in value, but during a growth phase its value might as well be zero. Whenever speculators believe the currency will actually have any long term value above its current exchange rate minus the resource costs, then it will be in a growth phase. The total resource consumption during this growth phase, including the supply chains to produce space, need to be studied. Measured in transactions per unit of CO2 emissions, I expect its efficiency to be better only during a phase of stable network space, i.e. when its value is not speculated to be higher than its current value.

To be fair you'd also have to measure and compare Chia's growth phase to Bitcoin's (or other PoW chains) growth phase as well. This would mean taking into account the manufacturing of the equipment, distribution and deployment costs of ASIC miners, electrical power generation costs, etc. Also, Chia runs on general computing hardware whereas ASICs have to be specifically designed and manufactured and ASICs use a whole lot more electricity than servers with hard drives running Chia. I think a single Ant miner uses something like 32 kWh a day IIRC. Chia has done a lot of power consumption analysis already and you can dive deep on that information yourself over at https://chiapower.org/ if you're curious. There's a dramatic difference between Bitcoin's PoW network and Chia's PoST network.