Feasibility of renting hash to process transactions

I am wondering how feasible it is cost wise to rent sufficient hash for approximately one block a day or week to consolidate transactions (lets say publish 1gb of transactions that accumulated).

It seems to a certain degree it is of advantage to be a smaller miner as long as large volumes can be offloaded each time the small miner mines a block (and there is no time sensitivity).