Warehouses are not uniform. If you were to pick any of the vast millions of acres of free land in the mountain west, you could simply pick a suitable flat area near existing power transmission lines and get to work. If you were trying to develop that capacity on warehouse rooftops you'll run into a variety of problems from: (1) Some warehouses do not have rooftops that lend themselves to power generation. (e.g. they include smokestacks, antennae, vents, and other features that make some of the roof unusable) (2) It is inherently more labor-intensive (and therefore expensive) to haul solar panels onto a rooftop and install them around existing infrastructure than to put them into place in a suitably-prepared empty spot on the ground. (3) Warehouses are not situated in spots that are optimally chosen for sunlight exposure and therefore may not be very productive for solar power. Many will exist in the shadows of mountains, buildings, or trees. Others exist in places where natural forces such as ice, hail, and rapid temperature swings will shorten the expected lifetimes of solar panels. (4) The inherent costs associated (labor, logistics, materials) with setting up 10,000 small solar plants and connecting them to the grid will naturally exceed the inherent costs of setting several large purpose-built locations and connecting them to the grid.