> Almost all of their grant money goes to paying the salaries of people in their lab (fixed costs).
That highly depends on the field. At my university in the math department, you're completely right. In the physics department however, the costs of the machines and experimental setups dwarf the salaries. Because the demand for things is so low, the cost is ridiculously high. For example suppose there were no plastic coke bottles, but you require one for an experiment. How much do you think it costs to make one coke bottle in a university's workshop, or to order one from a specialized company? The costs can easily go in the thousands of dollars. That's the situation scientists are in for parts used in their setups, and for the machines they use. If you can share machines, that can be a huge cost savings.
That highly depends on the field. At my university in the math department, you're completely right. In the physics department however, the costs of the machines and experimental setups dwarf the salaries. Because the demand for things is so low, the cost is ridiculously high. For example suppose there were no plastic coke bottles, but you require one for an experiment. How much do you think it costs to make one coke bottle in a university's workshop, or to order one from a specialized company? The costs can easily go in the thousands of dollars. That's the situation scientists are in for parts used in their setups, and for the machines they use. If you can share machines, that can be a huge cost savings.