Basic idea is to model demand with endogenous growth (but "satiation" becomes possible - eventually - at some notional "sufficiency" level); and supply then tracks demand with some time lag (~5-50 years - characteristic of commissioning/decommissioning large scale energy infrastructure). Then add cumulative pollution, with a hard constraint/limit which trumps demand and forces supply (of any non-zero polluting source) to zero. In the first instance we'll only have one source, and it will be polluting: so expect to see supply crash. Of course, "demand" will still carry merrily on its way up anyway, but the interpretation of the consequently growing supply shortfall will be left to the eye of the beholder...