> "Hydrogen economy" only makes sense if you have infinite free electricity or massive overproduction.

Or when batteries are really expensive and global production and/or geopolitics prevents a global power grid.

Both were the case 15 years ago (and geopolitics still prevents a global power grid today, but metal production has increased and is now sufficient).

Hydrogen wasn't entirely stupid back then; even though PV was more expensive than today, the trends were already clear.

Now? I think hydrogen is suboptimal for most users. But I wouldn't bet against the idea of someone, somewhere, likely in the arctic or antarctic circles, deciding that they really do need multiple months of energy storage, and for those specific weird edge cases I think it's at least possible they might decide a cryogenic liquid hydrogen tank the size of the space shuttle external tank, refuelled every summer by a comically large PV array that works 24 hours in some days, is less silly than 3 gigawatt-hours of batteries.

And don't you lose a significant amount compressing it for reasonable storage as well? Or is that considered part of the generation loss?

That's calculated in the total losses. You either need to compress it or freeze it. Usually for vehicles it's compressed, for long term storage or transfer it could be either.