Sharepoint has its origins in managing collections of MS Office documents more so than HTML and browsers. It knows about certain document types and tries to do intelligent things with them. It's not necessarily the tool you would use for serving raw data over HTTP with arbitrary Content-Types. (Given the complex and varied rules by which browsers interpret content, I'm not actually sure how one could even do that perfectly securely short of enforcing separate second level domain names for each and every tenant.)
As an old-school software engineer, we used to say the biggest part of requirements analysis is setting customer expectations correctly. It seems fair to say that the renaming of Sharepoint to "OneDrive for Business" has surprised some folks where it behaves differently from plain OneDrive or from raw BLOB store.