Some economists argue that DRS type LRAC curves don't really exist. They argue that such long run costs arise because of some "overlooked" fixed factor, meaning we'd actually be observing a short run cost curve. If these economists are right, we should never observe upward sloping LRAC curves.