It's a question of local vs global optimization. Do we only consider energy generation tech on its current path, or explore other possible paths to lower cost tech? It's unknown if some alternate energy tech could provide lower costs per KW unless they're explored. Even if prototypes cost more, one must try to at least imagine what the costs are at scale.

Beyond the direct costs there's also to cost of the new tech vs direct cost + external costs of current tech. One knows that burning natural gas has an externalities of adding carbon to the atmosphere.