> Even worse, the Bitcoin hashrate goes up as energy costs come down. The amount of energy burned is an equilibrium.
Sort of. What you say is true because the energy spent on mining is in equilibrium with the value of the Bitcoin produced by the miners (both the block reward and transaction fees). As the value of Bitcoin goes up, so does the energy they're prepared to buy to mine it.
As the block reward halves, the profitability of mining declines, and so does the energy miners are prepared to buy to mine it (otherwise they make losses).
The final equilibrium will be reached when the block reward goes to zero. At that point it'll be down to transaction fees only, which is proportional to transaction demand. If there's pressure there, then people will be incentivized to transact off-chain and use the chain for clearance only.
In the meantime, the energy spent (in fiat currency) is primarily a function of 1) block reward size; and 2) BTC value.
What's interesting about this is that the block reward schedule is fixed, but BTC value isn't. The energy "burned" is a consequence of Bitcoin's success - it's the area between the two curves. But had the block reward schedule tapered off faster, then Bitcoin might not have successfully bootstrapped at all.
The other interesting aspect is that the block reward, being inflationary, produces downward pressure on Bitcoin's value. It's the upward pressure (whatever its source) that is exceeding this that is creating the energy consumption (since energy consumption is proportional to BTC value, as above).