recently i been telling people.
like. look at it this way. before (what? when?) nvidia was a gpu maker and yearly revenue was maybe us$10b. and after. or now. its an accelerator maker and its revenue is nearer us$200b. im imagining the slurm/pbs market was, and still is, a fraction of that original us$10b. and the new us$190b are the kubernetes+mlops kind of thing. i wonder if theres a public breakdown of this somewhere...
and then i see nvidia bought slurm and then i see this https://slurm.schedmd.com/MISC25/Slinky-CUG2025.pdf

and yes the mention of hybrid clusters confounds me.
the way i see it is kubernetes shares cpu slices while slurm/pbs queues 100% cpu. ok not cpu, $resource.
they do absolutely opposite things. why are both involved , concurrently, in the same setup?