I think teams aren't lowballing free agents, rather, with the experience of the last decade, they're valuing free agents more accurately.
The presumption of a lot of the projections is zero injury risk, but on a ten year deal, injury risk is signifiicant, even for durable players at 26, much less players with any sort of injury risk approaching 30. So if you model as a decision tree, you have one outcome of 10 years at high performance with Probability Pa, a second won with 5 years of production then erratic production and a lost season or two with Probability Pb, and so on. You can use the universe of top players at their age 26 season to develop the probabilities of different outcomes.
A five year contract presents a much smaller risk and therefore merits a higher AAV.
For a team, it's not just expected value, but risk aversion, getting stuck with 5 out years at $30+M a year for 1-2 WAR per year is a disaster with a semi-firm salary cap.