It must be a limit, especially for container ships. The constraints can be the size of the existing berths and ports, the ports inside logistics and incoming logistics (container size cannot grow without changing the entire inland transportation system). What are the advantages of a larger ship? Less steel/mass of goods delivered, less fuel/mass of good delivered, lower labor cost/mass of goods delivered. Disadvantages are the longer time spent for uploading and downloading, less use for transportation, and increasing wall thickness. I think it is a nice optimization problem even without the port size and logistic issues, and neglecting the local productions and local demands.
In a recent report (October 2017) on the future of container ships, McKinsey&Company has assumed a maximum size of 50,000 TEU over the next 50 years compared to today's 22,000 TEUs.
The port's land, the cost and delivery time (transportation and charging /discharging) may be an inconvenient, however the optimization, analysis and conception can be nicely and smoothly done.