As someone that specializes a lot in data protection and archive solutions in all sizes of company, while technically feasible, this is a very bad idea, with a lot of risk in storing this data in this manner as described.
Tape would be an acceptable solution, as long as you wrote it to two copies of tapes. Then in 5yrs or so look at replacing your tape solution for a newer one. LTO is backwards compatible 2 versions back to read. e.g. LTO7 can read LTO5, and write to LTO6. So you may need to upgrade to LTO7 soon, to read those LTO5 tapes. Then in 5 or so years, go to LTO9, to be able to read LTO7 tapes. With a couple of tape migration projects inbetween. Or you have an enterprise tape library that can run multiple types of tape drives.
Another solution, would be to stick that data in the cloud, and keep it in two sites, or two providers. Like a copy in Amazon and/or azure. Amazon Glacier would be about $400-500/month in today's prices for 100TB. Their snowball can help you get the data into their cloud quick. Pricing should go down over time as scale and economics work in our favor. I suggest two vendors, because who knows how this new fangled "cloud," will shake up, and who wins and loses over the next decade.
If you are deadset on using hard drives, make multiple copies, and I'd probably do it across two different brands of drives.
When talking 10yrs of retention, the media type is very important, for compatibility sakes, and ease of accessing said data. But just as important is the environment, that this media is stored (humidity, temp, etc). Best to look at an Iron Mountain or similar to store these. Which by the way, I'm sure Iron Mountain offers some sort of storage platform, and long term archival solution too.