Because papers are the metric of a lab’s outputs, it makes sense for a lab to want any grant money they spend to be as “paper efficient” as possible. You can maximize your paper efficiency (ie. maximizing papers/$ or minimizing $/paper) in one of two ways. Either you just do cheap experiments or you spend the money on one piece of equipment that can be used for many many experiments. This drive may also be a partial explanation for the “salami slicing” phenomenon where researchers slice their research into absolutely minuscule pieces per paper.
While it is difficult to conclusively demonstrate (Counterfactuals are hard) the drive towards paper efficiency would suggest that there is a class of experiment that isn’t happening often. These experiments would take a lot of money per paper either in the form of just taking a lot of time or requiring a piece of equipment that would only be used for one or two experiments.
The lens of paper efficiency gives another explanation for the grad student explosion. There is an increasingly wide gap between the number of graduate students training for PhDs and academic positions for them to fill. Besides the fact that many grants have large carveouts explicitly for grad students, grad students themselves are the only thing you (as a PI) can spend money on that is self-motivated to generate papers. It’s conceivable that you could spend money on equipment that yields zero papers. Not the case with grad students!