Assuming they are akin to grad students, these "fresh graduates" likely have more intersectional knowledge than long-time deep domain experts. I've worked in grant funding landscape, and one of the biggest challenges is that:
1. Experts have deep knowledge, but don't keep up outside their field, and yet have the most power in distributing funds that bet on future innovation and 2. Grad students have very little power to direct funds, but straddle more fields and are better positioned to see new intersections emerge.
Each have strengths. Being naive has advantages to the learner that the expert can extract from simply by having them around
So I'd argue that keep an abundance of "fresh grads" isn't just cynical cost-cutting, but part of the secret sauce of why these big orgs generally operate as effective institutional ecologies