Actually, while the intention behind the proposal is kinda framed around enabling specific use cases without bloating the UTXO set, the actual technical impact on the network's overall resource consumption like storage, bandwidth, processing is still kind of a point of contention.
There are valid technical arguments on both sides of it and the long-term implications are not definitively settled yet , as I like to call it a 50/50 thing.
However I think removing the limit may actually introduce a couple of risks. While OP_RETURN data isn't part of the UTXO set, larger and more frequent data outputs will still increase the overall size of the Bitcoin blockchain and still demand more storage for full nodes and also increase initial synchronization time for new participants too.