Complexity IS difficult to do efficiently, that is true. I was simplifying. It is scaled by efficiency. Just because a competitor manages equivalent complexity with more efficiency doesn't mean the system is prevented from converging into a typical competitor dominance model. It just means the system is still transitioning. The complexity of the system increases because of it and the bar of entry for new competitors ratchets up. Competition decreases in proportion to how many competitors there are - which is strongly correlated to the cost of entry. The competition system (market) matures in an iterative process like this, with the overall effect of a gradual decrease in competition and an overall increase in the value-lead established competitors enjoy over new entries. It has the effect of the dominant position being occupied for a longer and longer time with a lower and lower probability of disruption until they are effectively perpetual monopolies.
If it is a market, information flowing from the consumer becomes increasingly decoupled and the actions of the dominant producer become less responsive. At that point, monopolies usually select a point that extracts maximum value from the system while retaining the minimal responsiveness required to suppress new disruptions as they arise. Generally they park there until they are dislodged by black swan events acting on the system from outside.
Mega corporations don't exist only because of governments. Governments tend to be just another example of this competition process. They will reform about as rapidly as monopolies once the system restarts and for identical reasons. I have been studying 'abstract competition' as a converse of 'abstract corruption' to look for features that may lead to alternatives, but as I said, it is a hard problem.
RE: Planned Obsolescence?