The key thing to remember about a non-redundant pool is that you lose the entire pool if ANY of the disks fail - so the more disks in the pool, the more likely it is to fail completely. A single striped pool of 98TB across 7 disks is a LOT of data to lose with 7x the likelihood of failure cf. a single drive.
My advice would be to have redundancy, and to create a single pool consisting of 2 RAIDZ1 vDevs - A) 3x22TB which gives you 44TB of space and B) 4x8TB which gives you another 24TB of space.
In reality, in total you would probably get about 40TiB-45TiB of useable space from this (bearing in mind all the overheads - such as TB to TiB conversion, metadata, and a realistic maximum utilisation of 80% etc.).
OTOH, with a striped pool, if a drive starts looking dodgy, you can remove it. Any data on bad sectors will still be affected, of course, but it may be possible to remove a failing drive before it fails completely.
I’m not one to recommend striped pools, but that’s what best meets OP’s stated objectives.