I don't like WAR at all, but IMO fWAR for pitchers is particularly bad.
Fangraphs uses what they think should have happened for a pitcher, rather than any measure of what actually did happen. They do this by using FIP. They only consider HRs, BBs, Ks, and HBPs.
Theoretically a pitcher could give up a base hit on every ball in play, but if there were no HRs, BBs, or HBPs fWAR would love him. If he managed to accumulate enough innings, he could lead the league in fWAR by giving up 6 hits and 3 runs per inning with 3 Ks mixed in.
bWAR for pitchers uses RA/9 as its base measurement.
For position players the primary difference between the two is what they use to measure defense.
IMO there are huge issues with both WARs. Here's a decent explanation of both.
https://www.samford.edu/sports-analytics/fans/2023/Sabermetrics-101-Understanding-the-Calculation-of-WAR#:~:text=Fangraphs allocates 570 WAR for,game compared to Baseball-Reference.