True.
Of course just because two people request one chunk does not mean that cache will automatically be used for the 2nd request.
The cache is likely to be between you and the vault in routing terms. So if you request a chunk 100,000 times then its very possible that 1 GET will occur and the caching occurs on your own machine or a node “close” to you. And another person asking for that chunk may cause a GET since the cache is “too far” away in routing space.