Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- Recalculate your congestion exposure as a share of the fleet, not a TEU count, before the October contracting round; a queue that grew 7.5% against a fleet that grew more than 7.5% is a smaller share, not a bigger one.
- Ask every carrier in the tender for the number of Suez transits it has run in the past 90 days and how many Asia-Europe loops it plans to return by 31 December; treat anything under 4 loops as unchanged.
- Price the Cape at the low end of the 5% to 7% band for a downside case and at the high end for a base case, and show the two rate outcomes side by side; the 2-point gap is a 33% relative swing on the midpoint.
- Convert the Asian congestion figure into days of dwell using your own gateway throughput, and use that number rather than the TEU total when you set safety stock for the fourth quarter.
- Cap any Asia-Europe contract signed in October at six months with a review on 15 January, and set the trigger at two additional loops returning to Suez before that date.
Linerlytica data shows Asian port congestion reached 4.3 million TEU, surpassing the 4.0 million stranded at the Covid peak, while Cape diversions still absorb 5-7% of global capacity. CMA CGM has made 15 Suez transits and Maersk 11, and MSC has returned its Jade service to the canal (calling King Abdullah, Singapore, Yantian, Ningbo, Shanghai and Qingdao), shortening Asia-Europe loops.
The Analysis
Four point three million TEU. The Covid peak was four million. That is the comparison in every headline this week, and it compares two numbers that were measured against two different fleets.
Linerlytica puts Asian port congestion at 4.3 million TEU, above the 4.0 million stranded at the peak of the pandemic. Cape of Good Hope diversions still absorb 5% to 7% of global capacity. CMA CGM has made 15 Suez transits, Maersk 11, and MSC has returned its Jade service to the canal, calling King Abdullah, Singapore, Yantian, Ningbo, Shanghai and Qingdao, which shortens the Asia-Europe loop.
Three claims, three different kinds of number, pointing in two different directions. That is the article.
Take the record first. Four point three against 4.0 is 300,000 TEU more, which is 7.5% above the previous high. As a quantity of boxes in a queue, that is a record. As a share of the fleet, it almost certainly is not, because the fleet in 2026 is larger than the fleet of 2021 and 2022 by a margin wider than 7.5%.
This is not a technicality. Congestion raises rates through the share of capacity it removes, not through the TEU it holds. If the fleet grew 15% while the queue grew 7.5%, the congestion share fell by roughly a sixteenth even as the headline set a record. Both statements can be true at once, and the one that prices your booking is the share.
The practical consequence is a false confirmation. A record TEU count reads as a capacity crisis, and a capacity crisis justifies whatever rate you were quoted. A falling share reads as a slow normalisation, and that reading is the one that should change what you sign.
Then the Cape. Five to seven percent of global capacity is a range, not an estimate, and the width of the range is the story. Two percentage points on a midpoint of six is a 33% relative uncertainty. On a low-end base of five it is 40%. A model that treats Cape diversions as a fixed input is carrying an error bar of a third.
Why the band is that wide says something about the measurement. Cape diversion capacity is not one object. It is extra transit days multiplied by the ships on the affected strings, and every input in that product moves: how many strings still go around, what speed they sail at, which ports they skip. A carrier that returns one service to Suez removes its share from the product. A carrier that slows a remaining Cape string down adds it back.
Which is why the Suez transits matter more than they look. Fifteen CMA CGM transits and eleven Maersk transits is 26 passages. Put that against the size of the thing it is supposed to reverse. One weekly Asia-Europe string makes 52 one-way passages a year in each direction, so about 104 canal transits for a single loop. Twenty-six transits is roughly a quarter of one string's annual passages, spread across two of the largest carriers in the world.
MSC's Jade service is one service with six ports on the rotation. It is a real change, because a loop that goes back through the canal turns faster and releases capacity. It is also one loop, and one loop does not reprice a trade.
The phrase the market is trading on, carriers return to Suez, describes a pilot rather than a pivot. That distinction matters because the two forces in this trade are pulling opposite ways at the same time. Asian congestion at a record absorbs capacity. Suez transits release it. The net is what prices, and anybody quoting either number alone is not quoting the net.
Split by trade direction, the two forces land on different people. An Asia-Europe exporter receives both: cargo sits in the Asian congestion, and the capacity released by Suez transits sits on the same lane, pressing his rate down. He is paying for a queue and collecting a discount at the same time, and the order of arrival is congestion first, discount second.
A transatlantic or intra-Asia shipper receives one: the congestion, without the release, because that released capacity does not come to his lane. His rate rises with the share and falls only when the queue clears. That asymmetry is why two shippers reading the same Linerlytica figure this week will reach opposite conclusions about whether to lock.
The sequence has a shape worth planning around. Congestion numbers are published weekly and read as a level. Canal transits are announced per service and take effect at a schedule change, which is six to eight weeks after the decision. Congestion data will therefore keep looking bad for a month after the capacity has already started coming back, and the market will price the bad print before it prices the schedule.
That lag is the window. If the Suez returns keep accumulating at this rate, the released capacity lands in the November and December schedules, and the congestion count will still be high when it does. Signing a twelve-month rate at that moment means paying for a stock that has already started to drain.
The part nobody is doing is converting 4.3 million TEU into a duration. A queue measured in TEU is a stock. What your booking costs is the dwell time it implies. Divide the Asian count by the throughput of the ports inside it and you get days, and days are the only unit that lets you compare this record with the 2021 record, because the ports are larger now too.
That brings the record claim back to its real weakness. Four point three million TEU beats four million TEU in absolute terms, across a fleet and a port system that have both grown. The comparable number is a share, and the share is the one figure in this story nobody published. When the most decision-relevant number is the one left unquoted, the absolute record is doing the work of a headline, not of an analysis.
Run the two scenarios. Suppose the Suez returns stop at 26 transits and the Cape stays at 7% of capacity. Congestion then absorbs a record share, the released capacity is a rounding error, and rates hold through the fourth quarter. Suppose two more loops come back in November and the Cape falls to 5%. Congestion share starts draining into a loose market and the rate you locked in October is the high of the cycle. The two outcomes are separated by two percentage points of capacity and by one decision at one network desk.
The trap is treating a record as a direction. A record is a photograph of a stock with no scale printed on it.
Lock the share, not the record. If you are signing an Asia-Europe contract in October, sign it for six months rather than twelve, because the two percentage points at the bottom of that range are worth more than the spread you are negotiating.
↑ Back to the key points