Global Port Congestion Deepens: Shanghai, Rotterdam, and Nhava Sheva Face Mounting Delays Into June

The Analysis

Put the whole container network on the table at once, not one port at a time, and look at the second week of June 2026. Shanghai is working through berth delays of two to seven days, the product of heavy fog and a cluster of vessels arriving together. Rotterdam and Antwerp, the two great northern European gateways, are seeing inland barge waits of up to four days. At Nhava Sheva, on India's west coast, vessels are waiting seven to eight days, because traffic rerouted away from the Middle East has compressed the capacity that used to absorb it. Three coasts, three very different problems, one week.

The temptation is to file those as three separate stories. They are not. They are one system, and the same mechanism is tightening all of them at once. The long detour around the Cape of Good Hope has been shuffling schedules for months, and when a schedule slips a vessel misses its berthing window at the origin, and that missed window does not stay at the origin; it travels, arriving at the destination as a gap someone else has to absorb. On top of that sits a container-equipment shortage worsening across the Asian export hubs, empty boxes stranded in the wrong places while the places that need them run dry. Put those two together and you have a network in which every node is being asked to be a little more perfect than it can be.

So this is not a note about fog, or rivers, or a single strait. It is a reading of a system discovering its own limits in several places at the same time, and of what an importer or exporter should do about it while the map is still moving underneath them.

Ten years ago you could read this network one port at a time and mostly get away with it. A delay in Shanghai was Shanghai's problem; a wait at Rotterdam was Rotterdam's. Each gateway had its own hinterland, its own fleet of feeders and its own rhythm, and when one of them set to a bad pace the shock had room to dissipate before the next node had to deal with it. That slack is gone. What has replaced it is a single circulatory system in which a blockage in one artery shows up as pressure two thousand miles away, and the 2 to 7 days at Shanghai, the 4 days at Rotterdam and Antwerp, and the 7 to 8 days at Nhava Sheva are not three coincidences.

They are three readings taken off the same instrument in the same week. When I say the real unit of analysis is the network and not the terminal, this is the week I would point to. Fog in one harbour, a river in another, a war-induced reroute in a third, and all of them landing on the same importer's delivery calendar as though a single hand had moved them.

The quiet engine underneath all of it is the Cape of Good Hope. When vessels take the long way around Africa instead of the Suez route, they do not merely spend extra days at sea. They break the discipline of the schedule. A ship that is running late into an Asian loading port arrives outside its assigned window, and berthing windows are not elastic; the box that missed its slot waits for the next one, and the next available opening is contested by every other vessel that is also drifting late. That is where the cascade begins, and it is why the report is careful to say that a missed berthing window at the origin propagates into further delay at the destination.

A cascading delay is a different animal from a delay. A normal delay is a cost you absorb once. A cascade is a cost that keeps being handed forward, node to node, gathering a little more at each step, until it reaches a consignee who had no part in starting it and no way to stop it. This is the chessboard logic of a network under stress: every late piece forces a move somewhere else, and the pressure you feel on your own square was created by a decision taken on the far side of the board.

A word on choke points, because the Cape is the clearest one the network has been taught this decade and I would not write about this week without it. A choke point is a narrow place on the map through which an unreasonable share of the world's trade has to pass all at once, and its power is out of all proportion to its size. A canal a hundred metres wide can set the tempo for a trade worth hundreds of billions, because when it is disrupted every vessel that would have used it has to find another way, and there is essentially one other way that does not add weeks. The Suez and the Cape are a seesaw, and the moment the balance tips, every schedule downstream of the pivot feels it.

What this week's numbers are really telling you is that the seesaw has been tilted for months, and the tilting does not stay at the pivot; it ripples out through loading windows in Asia, barge queues in Europe and berth waits in India as though the whole map were one trembling surface. Anyone who ships at scale should carry a mental picture of the choke points on their own routes, because a choke point that fails is the single fastest way for a quiet lane to turn into a crisis.

Now read the four numbers the way I would read them, because they are not all the same kind of number and knowing which is which tells you what to do. Shanghai's 2 to 7 days is a weather-plus-concentration problem: heavy fog reduces the rate at which a terminal can work, and vessels bunched up by upstream schedule slips arrive together, so the port is asked to digest a fortnight of arrivals in a week. Fog lifts and the range at the top of that band is not a permanent condition; it is the shadow of the pile-up. Rotterdam and Antwerp's four days is a hinterland problem: the ships may be fine, but the inland barge queue is where the gateway hands cargo to the continent, and a barge wait is a reminder that a port is only as good as its river.

Nhava Sheva's 7 to 8 days is a diversion problem: capacity that normally served one trade has been pressed into another by the rerouting out of the Middle East, so the vessels wait because the berths are doing more work than they were built to do. Three different causes, and yet all three express themselves the same way on your invoice, as days you did not plan for.

The equipment shortage deserves its own paragraph, because it is the part of this week that a rate sheet will never show you. Across the Asian export hubs, container equipment is getting tighter, which is a polite way of saying that empty boxes are in the wrong places. The box that should be in Ho Chi Minh City or Chattogram waiting to be filled is sitting at a port where nobody needs it, and the reason is the same schedule disorder that is stretching berth waits everywhere else.

Equipment is the network's circulation, the way blood moves through a body, and when the vessels run irregular the boxes run irregular with it. For an exporter this is the constraint that sneaks up behind the visible one. You can plan around a berth delay because you can see it coming; you cannot plan as easily around turning up at a yard to find there is no empty to load, especially when the empties your own incoming cargo will release are themselves stuck days behind schedule.

Equipment is worth one more sentence of the same treatment, because it is the constraint most likely to catch a well-prepared shipper off guard. You can plan to a band for berth waits and barge waits and still be stopped cold by an empty box that is not where you need it, and the shortage is not uniform; it moves with the schedules. The practical rule is to treat equipment availability as a separate, locally checked input rather than something you infer from the regional picture, and to confirm it at your own loading port before you commit a booking. In a week when three coasts are congested at once, the box is the quiet variable, and quiet variables are the ones that break plans.

There is a longer story here that is easy to lose inside a bad week, and it is the story of hubs themselves. The great gateways of any era are not permanent; they are the answer to a question the trade was asking at a particular time, and when the question changes the answer can change with it. A hub rises because it sits at a convenient point on the network of its day, and it can fade for the same reason, not because anyone decided against it but because the network quietly routed around it. What makes this decade interesting is that the questions are changing at once: where capacity can be absorbed, which routes are safe, where empty boxes can be relied on, which hinterland is worth reaching.

The ports that hold their place through a stretch like this are not necessarily the largest. They are the ones whose land, whose river and whose equipment balance can be relied on when the schedules slip, and that is a different list from the one a ranking table would give you. Watch the list of reliable nodes rather than the list of big ones, because in a network under stress reliability is the scarcer commodity.

Two more numbers complete the picture, and they belong to a different category, because they are forecasts rather than observations. NNR Global expects Shanghai's delays could reach 4 to 5 days during June, and reports that Port Klang is facing 4 to 6 day vessel delays on its Middle East and India routes. I would read a forecast differently from a reading. An observed 2 to 7 days is what the system is doing; a forecast 4 to 5 days is what a careful observer thinks the system will settle into, and the value of it is not the exact figure but the direction of travel. Both numbers point the same way: the delays are not a spike that will snap back next week. They are settling into a band. And a band is exactly the thing you can plan against, provided you accept it as a band and stop treating last year's tight transit as the norm.

Let me put a worked example on this, and let me be plain that only the ranges come from the report. The volumes, values and the cost of capital in the next few lines are my own assumptions, chosen so we have something concrete to hold, and you should swap in your own the moment you have your own numbers in front of you. Suppose you move 60 containers a month from an Asian export hub into northern Europe, through a gateway like Rotterdam. Suppose each container carries 50,000 US dollars of goods, and suppose your cost of capital is 8 per cent a year. Take the origin berth delay at the middle of the reported band, so 5 days, an assumption on my part rather than a published number, and add the 4 days of inland barge waiting that Rotterdam and Antwerp are reporting.

That is 9 days added to the door-to-door cycle before any cascade from a missed window is counted. Nine days spread across 60 containers a month works out to about 18 extra containers sitting in the pipeline at any moment, because 60 boxes times 9 days divided by 30 days is 18. At 50,000 dollars a box, that is 900,000 dollars of goods parked in transit that you did not budget to park, and at 8 per cent a year it costs roughly 72,000 dollars a year just to carry that extra inventory. And the calendar cost is the sharper one. If you had promised your buyer a 40-day lead time, and the network now adds 9 days of unavoidable delay plus the risk of more at the far end, you have quietly spent close to a quarter of your buffer on congestion alone. To promise the same date with confidence you would need to carry somewhere between 9 and 18 more days of cover on that lane, a decision that has a price whether you make it deliberately or let it be made for you.

A point is something you can hit; a band is something you have to live inside, and the shift from one to the other is the real change this month. For years the trade planned to points: a transit time here, a cut-off there, each one a single number everyone pretended was stable. What this report describes is a network that has stopped offering points and started offering bands, ranges two to seven days wide, four to six, seven to eight, and a range is not an inconvenience to be waited out, it is a fact to be designed around.

Planning to a band means you stop asking for the one true number and start asking for the distribution: how often the top end happens, what it costs you when it does, and what you will do on the days it does. A shipper who plans to a band carries a slightly larger buffer on purpose and sleeps better than one who plans to a point and is surprised every time the point moves. That is the mental switch this week is asking you to make, and it is worth making before the next report forces it.

So what do you do with a week like this. You stop planning to last year's transit times and you plan to the band, because the band is what the report is telling you is now normal. You take the reported ranges and turn them into a single internal number, your own planning transit, and you hold everyone to that rather than to the optimistic figure somebody still has in a spreadsheet. You treat the empty-container situation as a first-class constraint and get a written read from your forwarder on equipment availability at your specific loading port, not at the region in general, because the shortage is unevenly spread and the port that is dry is rarely the port you assumed.

You look at your routings and you ask, honestly, which of them have a single point of failure that a missed berthing window would turn into a week, and you give those lanes a second option that you have actually priced. And you build the cascade into your promise by committing to the far end of the band, not the near end, whenever a missed window would be expensive for you.

Every alternative carries a cost, and the useful thing is to know which cost you are choosing. Rerouting to a less congested gateway usually buys a shorter wait and hands you a longer inland leg, a fresh clearance process and a corridor you do not yet know. Switching the urgent portion of a shipment to air is a genuine circuit-breaker for a broken deadline, and an expensive habit if you let it become the default. Holding more buffer stock smooths the delivery to your customer while pinning down the working capital you would rather keep free, and the two effects pull in opposite directions exactly when money is tightest.

The trap in a week like this is not the delay; it is the delay you could have seen in a forecast and chose to ignore because it was inconvenient. And the specific trap of this particular week is the empty box, because it is invisible on the rate sheet and decisive at the yard gate, and by the time it bites it is too late to reroute around it.

I keep coming back to the same view of this thing. A network is not a set of ports, it is a set of promises, and the whole reason this week matters is that the promises are being made somewhere other than where they are kept. Fog over one harbour, silt in one river, a war in one region, a long detour around a continent, and a shortage of steel boxes three thousand miles from where they are needed, all landing on the same delivery date you wrote down on a quiet afternoon months ago. The sea will not organise itself around your calendar. All you can do is read the network as one thing, plan to the band it is settling into, and keep a second route in your pocket for the day the first one misses its window.

↑ Back to the key points

— By Tomás Ferreira