This obviously depends on local conditions. Where I lived then there was plenty of water so starting a large fire was not possible. The plenty of water also meant the ground wasn't stable. Of course they did need to maintain vegetation, but that is a known cost.
There's always another rat (or squirrel, or rabbit, or mole, or ...)
Growing up, this was something that occurred multiple times a year. Squirrels would climb the poles and short themselves across the 2 leads on the transformer. Zap, one less squirrel, and several hours without power while the power company came out to diagnose the problem and then send out for a replacement.
Independent vendors are faster, cheaper, and offer a more locally-adapted product.
The squirrels seemed to replace themselves without servicing. If only the transformers were capable of the same thing.
Throw in backhoes, erosion, etc. along with it being more expensive to fix when it does break, and overall it isn’t always guaranteed buried will be cheaper maintenance wise.
I don't know what PG&E will use in this case, but all buried power and transmission lines lines are robust against water infiltration. Solid-dielectric cables are rated for direct burial, and cooled lines are encased in steel conduit. Both types are sometimes further protected in concrete.
Not to mention that above ground lines have slack between poles while underground lines are pulled relatively taut. Any shifting in the ground is going to result in a break underground before a break above ground. Above ground lines are also better founded because the poles are sitting on foundations or driven deep so their ability to resist ground motion is a bit better while electric lines move with the ground around them pretty much in sync. This is particularly important in seismically active areas like CA but underground infrastructure actually typically performs better in actual seismic events.
While, yes, the cost to dig up a line is likely to be higher than the cost to repair a hanging line, that's just one cost and consideration. Economic loss due to wildfire as well as rolling brownouts to try and avoid wildfire are both costs you'd need to consider. You also have the cost to maintain vegetation around powerlines, vs basically no pre-emptive maintenance on buried lines (and little to no weathering).
And you, of course, have differing likelihoods of the line breaking in any case, which means the per-fix cost is only part of that picture. Lots of potential reasons, but I'd be curious to see a full study on those reasons for both above ground and below ground, their potential solutions, and their potential costs, but without that you have a very fuzzy picture.
"lo barato siempre es más caro"
The cheap is always more expensive.
see: https://en.m.wikipedia.org/wiki/Time-domain_reflectometer