But what would be the point? The miner wastes 500k on adding a bunch of, as you say, low value nodes to the network? What would they relay? Blocks with large op_returns?
Assuming BIP110 nodes are 10% of the network and the miner has added 10% to the total of relay nodes, the added nodes are basically doing what 90% of nodes were doing before. The added nodes would be even more pointless than the BIP110 nodes.
I would really like to understand your angle here and if this could be a threat to Bitcoin in some way.