Yes, 100% is possible. But having it happen doesn't mean much.
1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3
If you run this game through the analysis it will show 100% accuracy for both sides. All the moves are "book moves" which the analysis will give full credit for.
Now take the following game:
1. b4 1... g5 2. Bb2 f6 3. e4 h5 4. Be2 b6 5. Bxh5+ Rxh5 6. Qxh5#
It gives White an accuracy of 99.7, despite five "Best Moves" and one "Book". If I go to the Details tab, it gives White an accuracy of 99.68 on moves involving the queen (and also on moves involving the bishop or pawn.) But the only queen move was checkmate! So, for some reason, it's refusing to give 100% even on moves which are obviously best.
Don't make the mistake of taking those accuracy scores too seriously. The analysis runs only to a moderate depth and the results would change if it ran deeper (in your scenario with high-level engines playing against each other, the analysis would possibly mark some of the moves as not best just because it didn't see far enough to know why those engines played them.) And, clearly, there is something else going on which causes the analysis to not score moves as 100% even if those moves are perfect.