IMO the closest equivalent to an IPFS CID is a BitTorrent magnet URL. They both can be used to obtain an intermediate metadata value, which can in turn be used to fetch file content.

  > CIDs are still unique when referring to _metadata_, and can be used to
  > "unambiguously fetch content from its peers".

A CID, mathematically, cannot uniquely identify a single file. Consider the CID bafybcfbaqydfnrof5stdvkzhwx6kglb7p6owxja:

  [desktop]$ ipfs get bafybcfbaqydfnrof5stdvkzhwx6kglb7p6owxja -o temp.pdf
  [desktop]$ shasum -a 256 temp.pdf
  2bb787a73e37352f92383abe7e2902936d1059ad9f1ba6daaa9c1e58ee6970d0  temp.pdf

  [laptop]$ ipfs get bafybcfbaqydfnrof5stdvkzhwx6kglb7p6owxja -o temp.pdf
  [laptop]$ shasum -a 256 temp.pdf
  d4488775d29bdef7993367d541064dbdda50d383f89f0aa13a6ff2e0894ba5ff  temp.pdf

If you were to try fetching that CID from the peer network, which file you'd get would depend on which peer you're fetching it from.