One enduring insight from Brian Potter (Construction Physics) is that the cost of manufacturing tends to go down with scale, while the cost of construction doesn't drop nearly as much. Batteries are made in a factory and then pretty much just hooked up, while gas turbines have to be built. So, the difference in forecasts makes sense if you apply that rule.

There have been some experiments with prefabricated construction, but you still run into the issue that shipping really big things is hard, as we've seen with wind turbine blades. Robotic construction might make a dent at some point, but it still seems to be basically embryonic.