Furthermore, there's no interesting economic decisions being made by the player - power is simply something you must do to get full production. At best you have different types of power plants such as hydro plants in mountainous terrain, but that simply introduces an obvious optimum rather than a real decision to make. By contrast, I would point out that the current installation mechanics present a set of interesting decisions as part of the core gameplay loop, both decisions of kind (e.g., which kind of factory do I need? Which kind of mine? Which kind of shipyard?) and decisions of opportunity cost (e.g., dedicating more production to factories means less production is available for mines or shipyards). It is a simple yet effective set of mechanics leading to every player having their own tendencies and every new campaign having a different 'build order'.
In my view, if power plants will be added to the game it must be because they add interesting decisions to the core gameplay loop, not because they will make a nice, juicy target. I believe Steve's posts reflect a similar sentiment.
For interesting decisions, I think it'd be best to step back and look at planets as a whole.
Consider the planetary terrain. Does the player really care if the terrain is tundra, grassland, prairie, steppe, subarctic, cold steppe ? Not really. They have the same fortification and to-hit bonuses. There is no reason for the player to be concerned, as long as the planet is CC0.
Steve has made some moves towards interesting decisions on how to develop planets, with the upcoming mechanism where planets with terrain types affect the costs of training ground forces with terrain abilities - i.e. ground forces with "desert warfare" capability are cheaper to train on desert planets, than on grassland planets.
So lets look at the planetary terrain some more.
What would be the economic difference between a planet that is grassland, compared to a planet that is tundra ? My opinion would be that a grassland planet would have a much higher agricultural value, which would translate into fewer population required in the agricultural sector, and more people available as workforce for industry. Theory behind that is being able to farm on a greater scale, and less effort needed to maintain the farmland in a productive state - tundra requiring subsurface heating to allow ploughing etc.
Now that would mean that terraforming to
tundragrassland would be optimal for industry (for more workers). But would it ? Tundra planets are cooler, which means it's easier for thermal power plants such as coal/gas/oil/nuclear to cool their exhaust steam, making them more efficient. Which would translate into fewer thermal powerplants needed to power heavy industry.
Which means that despite a lower industrial workforce proportion on a tundra planet, the tundra planet might be a better choice for developing heavy industry on. Or developing particular heavy industries on. A grassland planet might be a better choice for light industry, and might have more production of trade goods (but not everyone plays with trade goods & civilian shipping on).
Another factor to consider for economic output would be the cost of planetary transportation. Mountainous, jungle, or rift valley planets might have higher transportation costs because of the difficulty in building railways. Planets with more hydrosphere might have lower transportation costs because more can be moved by (sea) ships, than by rails. Planets with dense atmospheres might have different transportation costs due to being able to build effective lighter-than-air freight airships.
Geological activity would also have an effect on a planets economy - it would favour geothermal power, and might make mining more productive by exposing new seams, but earthquakes and the like would make it more expensive to operate other industries.
So what does this all mean ?
It means that the player then has decisions to make about how to develop the planets they have available, decisions on what kind of biosphere they want to terraform a planet to, what industries they want to develop, and so on.
On the question of any additional infrastructure, I think it is useful to consider the effects that planetary governors have - appointing a different governor magically makes a planet 50% better at shipbuilding ? without any additional industrial construction necessary ? With that precedent, I don't know that implementing industrial infrastructure would be necessary. The effects of power plant efficiency and of planetary logistics could be modelled with just modifiers to outputs.
On the other hand, industrial infrastructure as something that can be built, is needed, and can be transported by civilian shipping lines, just like population infrastructure is, makes protection of shipping lines from raiders more of a priority. This gives the player more things to do, and disruption of enemy merchant shipping is then a higher priority, and allows a bit more in economic warfare.
So... with a bit more effects on population productivity (and morale ?) from terrain types, and effects on industrial productivities, there are then more decisions to be made on how to develop the planets, what industries to put on them, and how to develop them militarily.
Biosphere type - affects agricultural productivity i.e. the proportion of population required for agriculture - more productive planets have higher morale due to wider food diversity, but may be more vulnerable to radiation damage.
Biosphere type - also affects industrial productivity - the efficiency of power plants and transportation - jungles are harder to build things on, tundra have more efficient powerplants
Distance from star - affects industrial productivity - cheap solar power close to the star, more dependence on thermal plants further from star. Some biospheres would be more productive in some industries further from the star, whereas the same biosphere closer to the star would have more productivity from other industries.
Geological activity - affects industrial productivity - more maintenance due to earthquakes reduces most industry output, higher mining output due to minerals being more exposed
Industry types would have different productivities - Shipbuilding would be energy intensive, but not concerned with planetary transportation (Abstract - a shipyard would have plenty of surface-to-orbit transport facilities inherent in its nature, to move workforce and materials). Other manufacturing would have different needs - ordnance and fighter factories would probably require more transport due to their complex manufacturing processes (Abstract - a fighter factory is a network of smaller systems factories, engines, sensors, weapons etc, making transportation more of a concern), compared to construction factories (Abstract - construction factories are larger integrated facilities).
All these things could be represented by modifiers to output, rather than requiring infrastructure to be built.
And it might not be possible to optimise for everything simultaneously, thereby making decisions on how to develop planets more interesting.
