Sheesh, so what the hell is this thing? It's no regular hybrid/EV, that's for sure. Mining equipment? CAT, Komatsu, etc.? [Rhetorically. Not looking for a brand, and I suppose you'd rather not say who, which is fine.]
Some big bastard in any case.

Why ATF, why not coolant (~100C?)? Is the efficiency loss, due to the high operating temperature, really worth the extra what, 0.5% of conductor you're trying to squeeze in here? Is it already low-oxygen copper? How about silver? (Okay probably not, not at these scales...

) How about... GOSS or nanocrystalline core material, instead of regular silicon steel? (Alas, probably not feasible in motor stampings. Also the latter might end up as expensive as silver conductors would...)
Anyway, regarding insulation, I know there's some kind of silicone based enamel that goes over 200°C. There are power resistors with such stuff. It's hard and brittle, almost ceramic (and for that matter, probably decomposes and sinters into ceramic if you simply heat it high enough

), and probably not practical in the thin layers you're looking for.
Really, the amount of sliding, the potential for scraping and chafing, rules out a lot of things already. Like, even hard anodize aluminum (not that you can hard anodize copper anyway; heh, Al-clad then anodize?...ehhh...) isn't going to be all that reliable, shoving it into slots in silicon steel.
At least, that is, if assembly will be done in the usual way. If you have a different method in mind, like, the slots are slightly oversized, then after assembly, the laminations are crimped over the bars somehow, that might be cool. No idea. If you already have an assembly and inspection process pinned down, do disregard me.

There's also high temp epoxies, but I'm not sure how well they do quite that high, and give or take peak temps. You probably want something with a Tg above operating temp, so it retains strength and doesn't expand much. Shop around, see if you can find anything interesting, call up mfg reps. This is, after all, very much an industrial grade application, in need of an industrial grade solution; reps are still the way to go.

Also also, what processes are applied to the ends? -- I've seen these things before, they braze the bars to cross pieces to construct the turns. (At least, a local generator brand does it this way.) Which puts closer to 500°C on the bars, towards the ends of the slots. That'll damage even polyimide*, and decompose most silicones.
*Just barely. It can actually handle temps like that, momentarily (seconds?), but I'm guessing the bonding process will spend some minutes hot, so that kind of stinks. AFAIK, polyimide deteriorates by becoming more brittle, and probably carbonizing a bit, neither a great look for high voltage insulation. At least it would be after insertion force is applied, though.
If nothing else, I would definitely feel more comfortable if you can afford the space to fit fiberglass or mica paper wrapping, and an impregnation step doing something with a high temp resin, with probably a high mineral fill content* as well.
*Incidentally, semiconductor packaging for example is something like 80-90% silica fume, which is how it gets a higher CTI, and such low expansion rate. Not quite directly applicable here I think, as the resins top out around Tc(max), 150-175°C, so, not much over 200. They can sustain soldering temps though. Maybe someone makes an even higher temp version?
Tim