Most large modern buildings have active ventilation built in, meaning that electric fans drive the air through the building. The airflow direction is usually fixed at construction time. However, if the wind happens to blow from the opposite direction to the ventilation flow, then the fans require extra energy to counter the wind. On the other hand, if the wind agrees with the airflow in the building, then the fans may not need to be run at all. To save electricity, a building could have a wind direction sensor (a weather vane) on the roof connected to a switch that reverses the ventilation fans, so that the fans always pump air in approximately the same direction as the wind. If the wind is strong enough, a wind speed sensor (a small windmill or windsock) on the roof could stop the ventilation fans altogether.
The tradeoff for this adaptive ventilation system is the initial fixed construction cost and the ongoing maintenance of the weather vane, windsock and controller of the fans. All the extra components of the system (relative to the current unidirectional ventilation) are cheap and robust, so the both the fixed cost and the maintenance should be negligible.
Current ventilation systems have differently shaped air inlets and outlets in the rooms, which suggests that the system requires a particular airflow direction. In this case, adaptive ventilation may be much more expensive than the current ones, because the ventilation shafts and air vents need to be doubled. To avoid the need to build twice as many shafts and vents, have just the air inlets and outlets of the whole building switch roles with the wind direction. The rest of the system can remain unidirectional when the valves from the building’s inlet and outlet to the rest of the system switch appropriately. The air inside the building can then move in the opposite direction of the wind some of the time. In this case, the electricity saving is only realised if the building is sufficiently airtight, which is the case for modern highrises that have unopenable windows. If the air is allowed to move through the building independently of the ventilation and the wind is opposite the airflow in the system, then the fans have to overcome the air pressure difference like in the current systems. This wastes electricity.
My idea of a ventilation system in an average modern large building (which comes mostly from movies/games, so please correct me if I’m wrong) is that it has a cluster of intakes on the roof, and its outputs are presumably also clustered there.
Given this structure, it probably makes more sense to make these inputs and outputs rotateable than making the whole system able to work both ways. Design and maintenance costs of one adaptive mechanism must still be an order of magnitude below those for a thousand similar smaller mechanisms designed for each room.
The second option is making both inputs and outputs static but universal — i.e., able to work with any wind direction. Think orthogonal to wind. Vertical. Pipes. One pipe sucks air in, another one lets it out. That’s a risk-averse solution.
Why was this option not used by architects who were presumably aware of it (pipes have been around for a while)? Likely answer: wind rose. Wind is not completely random, and at any given place it more often than not blows one way more frequently than the opposite. Other tall buildings in the area further shape wind flows (though uncertainty about future building plans in the area may add to the unpredictability). All this statistics allows to shape even static air intakes and outputs optimally, meaning that gains from making them dynamic are also smaller than if the wind was uniformly random.
The rotateable intakes and outlets are a clever idea. These may take less or more energy to rotate than reversing the direction of the ventilation fans, because for the fans, the wind helps reverse the direction. To get the wind to help rotate the ends of ventilation pipes, weathervane-like fins could be attached to these which always point the intakes towards the wind and the outlets away from it.
The gains of reversing the airflow are indeed the smaller the more predictable the wind direction is. However, in most parts of the world the wind direction varies quite a bit (the wind rose may not be uniform, but is far from a unidirectional spike). Other buildings make the wind even less predictable and cities are not static. The construction of new highrises nearby will change the wind pattern for a given roof.