I'm reading about SegWit and the BIP 141 in particular and I don't understand what is the motivation behind the P2SH nesting.
At first I thought the P2WPKH is not valid on non-upgraded nodes, so P2WPKH nested in P2SH is a way to make this a soft fork (with the P2WPKH being used only after all nodes are upgraded). But as I was looking into the definition of P2WPKH I came to a conclusion it is a ANYONE_CAN_SPEND transaction on the older nodes.
My another guess was that this might be to enable a segwit-unaware wallet to create a transaction containing a scriptPubKey, which can be redeemed later on using a segwit witness. Is that right? Are there any other benefits? It seems like a little bit too much of a hassle.