Yantian and Shekou reopen but appointment shortages replace typhoon closures

The Analysis

Put the Pearl River Delta on the table and trace the coast from Yantian up the eastern shore of Shenzhen Bay, then west and around to Shekou. Typhoon Noul has passed. On 27 July 2026 both terminals resumed operations, and by breakfast on that same day every gate appointment at both facilities was gone. The weather cleared; the queue did not.

Anyone who has run this corridor for a few years knows the shape of the thing by now. A storm can shut a gateway in a night. The recovery takes far longer than the storm, because the storm only stops the water, while the recovery has to pass through a gate, and a gate is a much narrower thing than an ocean. Congestion has replaced weather as the binding constraint, and it will still be the constraint long after the sunshine is forgotten.

I look at the Pearl River Delta the way I look at any hub: as a piece of plumbing with a mood. Two terminals, a bay between them, feeding the same hinterland of factories, the same fleet of trucks, the same off-dock yards that were never designed to hold a week of a port's output. When both terminals are down for the same storm and both come back on the same morning, the corridor does not simply switch on. It inhales. That intake is what the next two weeks are about, and the gate rules announced on both sides of the bay are what will decide how it goes down.

Ten years ago this corridor had a different temperament. The terminals were busy, but they were not yet the spine of so much of what leaves southern China, and the land and the roads behind them had more slack in them. A closed gate used to be an annoyance that cost a shipper a couple of days. Now every hour at a gate matters, because the factories behind it run on schedules that have been squeezing inventory out for a decade, and a truck that cannot get in is not just a truck waiting; it is a production plan with one less day of cover.

Both terminals resumed on 27 July, and that sentence is the good news. The bad news is in the second half of the same report: the day both reopened, the gate appointments were fully booked, the roads around Yantian were reported severely congested, and yard pressure was described as running high. The constraint had moved from the sky to the gate, and a gate does not clear the way a sky does. A closed sky opens all at once; a gate opens one slot at a time, and only as fast as the yard behind it can be emptied.

Here are the mechanics, and they matter more than the headline. Yantian has restricted export full containers so that they can only gate in within seven days of a vessel's estimated time of berthing, which the trade writes as ETB-7, and it has capped daily intake at roughly 9,000 boxes. Shekou will adopt the same ETB-7 rule from 3 August, and for each service it will set a ceiling of 120 per cent of the maximum single-sailing loading over the past three months. Meanwhile the off-dock yards have accumulated heavy backlogs, and the costs of waiting have started to move: storage charges and truck waiting charges are rising, and the risk that a box misses its vessel cut-off is climbing.

Read the afternoon of 27 July through a shipper's eyes and it says one thing. The storm stopped the water for a couple of days; the recovery is now being rationed through a gate, and the gate has rules it did not have before the storm, because a backlog admitted without discipline would simply turn into a yard that cannot work. There is a geography to this, too. Yantian and Shekou sit on opposite shoulders of the same bay, roughly forty kilometres apart by road, and they pull from the same pool of trucks and the same band of off-dock yards. When both cap their intake on the same day, it is not two terminals that get tight. It is one labour market for drivers, one supply of yard space, and one road network that suddenly has to absorb the orders that both gates are turning away.

Think of the gate as a valve on a system that has just been pressurised. For years the Pearl River Delta ran as a fairly forgiving loop: a box arrived, sat in a yard, found a truck, found a vessel, left. The variables were loose enough that no single one of them could stop the flow. A storm changes that, because it fills the whole system at once and then hands control back to the narrowest point, which on 27 July was not the berth and not the vessel but the gate and the road outside it.

ETB-7 is the terminal's answer to that pressure. By refusing full export containers that arrive more than seven days before their sailing, the terminal stops its yard from filling with boxes that have nowhere to go yet, and by capping daily intake at about 9,000 boxes it keeps the road outside from turning into a permanent car park. Both policies trade a shipper's freedom to deliver early for the corridor's ability to keep moving. You can argue about whether that is the right trade on a quiet Tuesday. On 27 July it was not really a choice.

Step back from the fortnight and the delta's role is not hard to read. This is the corridor where a decade of electronics assembly and cross-border e-commerce has concentrated, and its two main shoulders have acted as the valves for a hinterland that reaches far past Shenzhen. A rule like ETB-7 is not really about one typhoon. It is the terminal telling the market that the corridor has reached the point where discipline at the gate buys more capacity than concrete does, and that is a message you will hear again at every busy gateway in Asia over the next few years.

The gate, not the berth, is where a modern hub is actually constrained, because you can berth a bigger ship within a decade but you cannot widen the road outside a terminal in a season. Anyone who learns to plan around the gate now is a step ahead of the ones who are still planning around the vessel. Seen from the water, ETB-7 looks like a small administrative change. Seen from the road, it is a different port. When a terminal decides that a full export container may not enter until seven days before its sailing, it has redrawn the boundary of where a port's operation begins: the queue outside the gate is now part of the terminal, whether or not the terminal carries a line item for it. That boundary shift is the real news of 27 July, and it will outlast the backlog that produced it.

There is one more piece of the machine that never appears in a terminal's own description, and it is the trucks. A gate cap is not just a limit on boxes; it is a limit on how many times a tractor can turn in a day, and a tractor that waits three hours outside a gate instead of ninety minutes does not make that time up later in the week, because a driver has a legal day and a home to go to. The corridor's truck fleet is finite and shared, and when Yantian and Shekou both ration intake, the same fleet is being asked to serve both shoulders off a road network already reported as severely congested.

This is why the cost of a gate rule shows up first in truck waiting charges and only later in everything else. It is also why the fastest thing you can do for your own cargo is not to negotiate with anyone but to move your boxes into the early part of their window and keep the tractors turning. A box gated in on day three is not only cleared; it frees a tractor to carry someone else's box, and in a rationed week that second-order effect is the entire game.

Let me make that concrete, with the assumptions stated, because none of these numbers except the caps and the seven days come from the report. Assume your own book carries 500 export containers that need to gate in over the next ten days. Assume, further, that the corridor as a whole is carrying a post-typhoon backlog of around 60,000 boxes. That 60,000 is my illustration, not a published figure, but it is the right order of magnitude for a two-terminal corridor coming back from a two-day closure. At Yantian's stated ceiling of about 9,000 boxes a day, 60,000 boxes is seven days of gate throughput if the queue behaved perfectly and every box at the front was ready to load.

It will not behave perfectly, and not every box at the front is destined for the next sailing. Your 500 boxes sit somewhere inside that queue, and their order is set not by when they reach the gate but by vessel cut-off: a box for a ship sailing on Thursday overtakes a box for a ship sailing a week later, every time. That is what ETB-7 formalises. The terminal keeps a clean seven-day window and pushes everything outside it back into the yards, which is exactly where your boxes are, and exactly why the storage and truck-waiting meters are running. On the Shekou side the same logic wears a sharper edge.

Assume a service whose recent peak sailing lifted 8,000 forty-foot units. From 3 August its ceiling is 120 per cent of that, so 9,600 units per sailing. On a normal week that ceiling is invisible. In the week after a typhoon, when every shipper in the corridor wants the same two or three departures, the 1,600 units above normal lift are the entire surge capacity the service will take, and they will go to whoever is gated in and documented first. Put a truck number next to the box number, because that is where congestion actually bites. Assume a haulier's tractor makes three gate turns a day on a good day and two and a half when the approach road is clogged; over a ten-day window that is five turns lost per tractor, and if twenty tractors are carrying your 500 boxes, the corridor is short a hundred turns, which is a hundred boxes' worth of capacity that does not exist and cannot be bought back. You cannot order more trucks into a jam; you can only cut the number of times each truck needs to enter, by consolidating loads and gating early. That is the arithmetic behind pushing boxes into days three to five of their window.

Today is 27 July, so the useful question is what you do this week and next. Name a corridor owner by 29 July, and make it the person who owns the export cut-off calendar rather than the forwarder's account manager, because this is a scheduling problem with commercial consequences and it needs a single throat to choke. By 29 July you want one list of every export container you have within ten days of sailing, sorted by cut-off, with each ETB-7 window shown as a date and not as a rule. Everything sitting outside its seven-day window is in the off-dock yards by definition, so those are the boxes where storage and truck waiting are already accruing, and those are the boxes whose cost of waiting you should track daily from 28 July and put in front of the finance owner, not keep buried in the forwarder's monthly statement.

Set a trigger and write it down: any box that is more than five days away from its ETB-7 window on 29 July gets a decision that same day, either a re-plan onto a later sailing or a deliberate early delivery if the terminal will accept it. Do not let boxes drift toward the window and then discover the problem at the gate. For the week of 3 August, when Shekou joins ETB-7, assume both shoulders of the bay are rationing capacity at the same time, and build your slot bookings two days earlier than usual rather than one. Set a goal you can check on 7 August: no more than 10 per cent of your in-window boxes gated in on the last permissible day of their window, because a corridor where everyone gates in on day seven is a corridor that clogs on day seven.

Every box you shift forward into days three to five of its window shortens the queue behind it, and against a fixed daily cap that is not a courtesy, it is throughput. The same discipline applies to trucks. Instruct your hauliers on 28 July that a booking is confirmed only when it carries a date inside the ETB-7 window, so nobody drives to a gate on a hope. Put a number on the haulage side as well: by 30 July, ask each haulier how many tractors it can commit to the corridor for the week of 3 August, and treat that number, not their reassurance, as your real capacity. If thirty tractors can give you a hundred and twenty turns a day and your in-window boxes need a hundred and fifty, the shortfall shows up on paper two days before it shows up in a yard, and paper is the only place a rationed week ever lets you see ahead.

Two alternatives are worth having in your pocket, and both carry a price the terminals will not pay. Route part of the book through a second gateway in the delta, or through a neighbouring region's port if the lead time allows, and you buy relief from the Yantian and Shekou caps, at the cost of a new set of gate rules, a new off-dock yard and a truck leg you may not have rehearsed; an untested gateway is a rumour, not a route. Or hold the cargo deliberately and accept a later sailing, which costs you the customer promise rather than the storage bill, and is the right answer only for product that can genuinely wait.

The pitfalls cluster around three things. The first is treating ETB-7 as a formality and building the truck plan on when the gate is open rather than on when your booking window closes; the two are now different, and the gap between them is where boxes get stranded. The second is documents: a box re-planned onto a later sailing, or moved to a different gateway, needs its paperwork to move with it, and the moment a re-planned container's documents lag behind the cargo it sits at the gate while the terminal and the customs record disagree. The third is the off-dock yard itself, because a yard that is overflowing does not only cost money, it loses boxes, and a lost box on the wrong side of a cut-off is not a charge on the invoice, it is a conversation with a customer.

One quieter trap deserves naming: assuming the surge clears in a week. A two-day closure in a corridor running close to capacity takes longer to unwind than the closure itself lasted, and with both shoulders of the bay going through it at once, the unwind is measured in weeks, not days.

A port is a strange kind of machine, because it only fails visibly. When it works, nobody sees it; a container simply appears where it was meant to, and the buyer never learns the name of the terminal it left from. When the wind comes and the gates fill, everyone in the chain suddenly knows exactly which yard, which road, which cut-off, because those are the names on the emails at three in the morning. Yantian and Shekou are open again, and on any long view that is the whole story: this is a delta that has swallowed worse weather and kept moving, and the boxes will move.

But the long view does nothing for the 500 boxes in an off-dock yard that are running out of window right now. The one thing worth watching this fortnight is not whether the terminals are working. They are. It is whether the corridor learns to breathe in with some discipline, and how many weeks it takes to breathe out. The sea came back the day the typhoon left. The land is still catching up, and the land is where your boxes actually stand. I have watched a lot of hubs come back from a lot of storms, and the ones that recover cleanly are never the ones with the most cranes.

They are the ones where the shippers stopped treating the gate as an afterthought and started treating it as the part of the port that decides everything. Yantian and Shekou will be forgotten by the shipping pages within a week. Your cut-off calendar will remember them for a month. Keep your eyes on the month, not the week.

↑ Back to the key points

— By Tomás Ferreira