But it consumes time and effort to hire/rent people and creates unnecessary churn to have people all running at 100%. The machine version of this is nicely discussed in a book written by Eli Goldratt called The Goal with a more people & project based take on it called Critical Chain. They're in novel form (there are more serious treatments available of the general Theory of Constraints), but it discusses why running each resource, person or machine, at 100% in an organization isn't actually the most efficient thing to do for an organization.

https://en.wikipedia.org/wiki/Eliyahu_M._Goldratt

What often happens is consultants/temps are leased to fill in temporary gaps. Short term it's cheaper (and faster) than hiring for permanent positions.

And sure, people typically don't run on 100% capacity. What however seems to happen in many bigger orgs with vague performance metrics (as much of IT work is), some people do solid 60-80% productive work throughout the day, while some barely hover over 0%. So you have a kernel of workers pushing the cart and a bigger cohort of watercooler crowd. When peak load time comes the latter don't ramp up their productivity to 100%, but instead it's expected the employer should expand the staff.

Nothing short of impending layoff rounds seem to affect this dynamic.