What will get very interesting, in my opinion, is when WPO with nanite gets finished. Truth be told, I’m not entirely sure how well it would work, as needing to do per-vertex deformation gets expensive fast with large instance counts. Still, optimism can’t hurt.