In the real mining industry we see the same effect, where there is a commodity price (for say Nickel or Iron), and the suppliers, faced with very high capital costs, will continue to produce until their marginal cost is underneath the marginal revenue. At this stage they slow or mothball mines, smelters and refineries. That comes at a huge cost.
The last mines left are the ones with the most efficient costs/tonne, which is combination of access to ore, the type of ore, the closeness and efficiency of smelters and refineries and so on. A big part of it is the scale of the operation, and how up to date the equipment is, but also how well it is all operated.
Aluminium refining is particularly relevant, as electricity is one of the main input costs. We have seen refineries close down and sell their power back during extended periods of high electricity prices and relatively low aluminium prices (e.g. when Enron dove up Ca. prices). Newer/larger aluminium plants are fr more efficient, but the price of electricity is everything.
(I've led turnarounds at very large nickel and alumninium smelters/refineries and helped others with their turnarounds on mines/operations)
In the case of bacon mining the main factors are, once the computing equipment is acquired, the price of electricity and the conversion efficiency per kWh.