On October 14, BIMCO reported, citing Clarksons Research, that container ship capacity transiting the Suez Canal averaged 753,000 TEU a week in the first 2 weeks of October, the first time since the Red Sea attacks began that transits have passed half of the 2023 level. The Suez Canal Authority's own figures, reported by The Maritime Executive on October 13, show 166 large container vessels transiting in the third quarter against 16 a year earlier. The same article quoted Drewry's estimate that the direct saving from returning is about $43 per slot, roughly $1 million per round trip before war risk premiums.
$1 million per round trip sounds like a lot until it is set against the revenue of a 20,000 TEU vessel. It is not the number that moves a carrier. The number that moves a carrier is ships.
A loop is a fleet, not a voyage
A weekly liner service is a fixed frequency loop. The number of ships it needs is the round trip duration divided by 7. Take an illustrative Asia to North Europe loop with an 84 day round trip via Suez: 12 ships. Route it around the Cape and add roughly 14 days each way: 112 days, 16 ships. The 4 extra ships are not a cost line on a voyage estimate. They are a third more fleet on that string, chartered or diverted from somewhere else, for as long as the diversion lasts.
Reverse the diversion and those 4 ships come back. That is the real economics of the return: not a toll saving per slot but 4 vessels per string released into a market where they can be redeployed, laid up, or used to absorb the newbuilds still arriving. Multiply across every Asia to Europe string a carrier runs and the decision is about fleet size, not fuel.
Why the return is slower than the exit
The exit was fast because it was forced: a vessel at risk reroutes within a day. The return is slow because it is a network change that only pays once the whole string moves. 1 ship going through Suez while its 11 siblings go round the Cape wrecks the schedule interval and saves nothing. The string has to be rephased as a unit, which means a gap of several weeks in which the loop runs short or doubles up, and customers see blank sailings or bunched arrivals. The cost of switching is paid up front and the benefit is collected over the following year.
That asymmetry explains why carriers returned 1 string at a time, and why the figure passed 50 percent nearly 3 years after the diversions began rather than a few months after the attacks eased.
What a planner should take from it
Route choice on a liner network is a fleet allocation problem with a switching cost, and it should be modeled as one. The variable to optimize is not voyage cost; it is the number of vessels committed per string under each routing, weighed against the probability that the route has to be abandoned again. A model that prices the return on fuel and tolls alone will always say go. A model that prices the rephasing gap and the risk of a second exit will say go only when the expected time on the route is long enough to pay back the switch.
The carriers that waited were not slow. They were pricing the second number.