Transpacific Peak Refuses to Break: Ocean Alliance Loops Clear $10,000 a Box

The Analysis

"Standing on the buying side of the desk, the talk at the 140th Canton Fair did not sound like a market about to cool." It sounded like a market that has stopped believing it will. Every forwarder on the floor was telling the same story: the peak that was supposed to break in October has not broken, and nobody on the ground is booking on the assumption that it will.

The numbers back them, and they are ugly in a particular way. Drewry put Shanghai to Los Angeles at 8,007 dollars per feu on 22 October against 2,290 dollars for the same week a year ago, up 250 per cent, and Shanghai to New York at 10,166 dollars, up 197 per cent. United States containerised imports reached 2.55m TEU in September, up 10.3 per cent on the year. Congestion is absorbing roughly 12 per cent of vessel capacity, mainline operators are omitting Chinese load ports altogether, and the squeeze is expected to run through to Chinese New Year.

None of those figures is the one that should worry a buyer. The one that matters is buried in the middle of them: Ocean Alliance ordinary loops, quoted at around 4,000 dollars before March, have gone through 10,000 dollars, above Matson's premium service. Ordinary has overtaken premium. I do not distrust that number because it is high. I distrust it because it tells me the thing I have been paying extra for has quietly stopped being sold.

Lay the fact slice out flat before arguing with it. Drewry's 22 October assessment had Shanghai to Los Angeles at 8,007 dollars per feu against 2,290 dollars in the comparable week of 2025, a rise of 250 per cent, and Shanghai to New York at 10,166 dollars, up 197 per cent. On the demand side, United States imports reached 2.55m TEU in September, up 10.3 per cent year on year. On the supply side, congestion is taking roughly 12 per cent of vessel capacity out of the trade, and the mainline operators are not merely blanking sailings, they are omitting Chinese load ports altogether. The horizon everyone at the fair was working to is Chinese New Year, which in a normal season is where this ends.

Why has it not broken? Start on the demand side, because that is where the surprise sits. Imports up 10.3 per cent in September is not a panic number, it is a steady one, and steady demand is harder to wait out than a spike. A spike you can sit through. A run of months at plus ten per cent means the inventory drawn down over the summer has to be rebuilt, and rebuilding runs on a clock set by shelf space and sales forecasts, not by freight rates. Buyers cannot postpone their way out of that.

Now the supply side, where the arithmetic turns nasty. Congestion absorbing roughly 12 per cent of vessel capacity does not remove 12 per cent of the boxes, it removes 12 per cent of the space, and then it removes more. A ship that waits three days for a berth does not sail its next rotation on time, round trips lengthen, the same hulls deliver fewer annual sailings, and the effective loss compounds. Layer on top the omitted load ports, which is a quieter version of the same withdrawal: your booking is not cancelled, it is rolled, and a booking rolled in October is an empty shelf in December.

Put the two sides together and the inversion stops looking strange. Matson's premium is a price on time: fewer days on the water, a more predictable arrival, sold under contract. When congestion stretches everyone's transit by the same week, days stop being something you can buy, and the premium has nothing left to sell at the exact moment the market is shortest. On the ordinary side, the anchor has moved. Those loops are no longer priced on how fast they are, they are priced on whether your container gets loaded at all.

You can see the mechanism in the way space is being sold rather than in the rate itself. When a loop clears above a premium product, what the market is rationing is not transit, it is the right to load at all. Listen to what your forwarder asks you for this month. If the conversation has moved away from price tiers and towards allocation, minimum quantities and penalties for booking late, you are no longer buying transport, you are buying a place in a line, and the two are priced by entirely different rules.

Who bleeds first. The spot-exposed shipper with no contract and no allocation bleeds in the current week, and bleeds twice: once on the rate and once on the roll-over, because a booking without a loading guarantee is a booking that can be moved. Small and mid-sized importers live here, and so does anyone who let a contract lapse rather than renew into a number they did not like. Then the shippers on quarterly or short-term agreements, who meet the increase at the next reset and discover their rate was indexed to something that moved without them.

Who bleeds later, and for more. The buyers whose renewal falls in the first quarter. Their next twelve months get priced against a spot rate near 8,000 dollars rather than the 2,290 dollars that set this year's baseline, and they carry that for a full contract term. That group is larger than it looks from inside any single company. And who does not bleed: cargo still moving under a fixed annual rate that has not expired; non-vessel-operating common carriers holding fixed carrier contracts, whose spread widens for as long as their buy and their sell have not both reset; and the Chinese exporter selling FOB, who does not pay ocean freight at all.

The non-vessel-operator sitting on a fixed carrier contract is the quiet winner in this, and it is worth understanding why, because it shows you where the margin has gone. If his buy has not reset and his sell has, the spread belongs to him. If both have reset, he is in the same position as everyone else and simply passing it through. Find out which of the two you are dealing with before you assume his quote is structurally cheaper.

That last one deserves its own paragraph, because it is where a good many exporters misread their own exposure. On FOB terms the overseas buyer carries the freight, so the increase lands on their account rather than yours. But your buyer's landed cost has still risen by double digits, and a buyer whose landed cost jumps does three things without being asked: postpones orders, opens a conversation about the ex-works price, and starts costing an alternative source nearer to home. Not one of those appears as freight in your books. All of them are freight arriving by another door, and by the time you recognise them there is a purchase order attached.

When does it reach your account? The spot rate reaches you this week, in the quote sitting in your inbox. The contract rate reaches you at the reset, and that is the date worth working back from. Shanghai to Los Angeles runs roughly two to three weeks on the water, so cargo booked in late October lands in November, and anything meant to govern the first quarter has to be papered in November and December. Two renewal windows matter in this trade, 1 January and 1 May, and the January signatures are the ones being priced against today's spot.

Then the horizon, and it is not the clean date it appears to be. Chinese New Year is not a single day on a freight calendar. Factories wind down ahead of it, the loading rush builds for weeks before that, and the restart after the holiday takes another few weeks to settle. Sign in December to cover February arrivals and you have signed into the worst fortnight of the year. Price that rather than meet it. And be clear about who is insulated and for how long. A buyer still sitting inside a fixed annual rate is not paying today's spot, but he is paying for the renewal that comes out of it, and that renewal is negotiated in the window when spot is at its highest. Insulation and immunity are not the same thing, and the distance between them is one signature date.

Here is the part the coverage has skipped, and it is the reason this is worth your time. Every write-up has treated the inversion as a curiosity: ordinary loops dearer than the premium service, how odd. It is not odd. It is a product failure, and you are the one standing in the shop holding the broken product.

A premium service sells time. That is the whole of what the extra money buys, fewer days on the water and a more predictable arrival. Congestion has taken roughly 12 per cent of vessel capacity out of the trade and lengthened round trips across the board, which drags the premium transit and the ordinary transit towards each other. As the gap in days closes, the premium has nothing left to sell. On the ordinary side the anchor has changed: those loops are being priced on whether your box gets loaded, not on how fast it moves. That is why ordinary crossed above premium, and that is why the crossing is a signal rather than a curiosity.

Which means most buyers are aiming at the wrong variable. They walk into these renewals arguing dollars per feu, and dollars per feu is not the variable that decides anything in a market that allocates by queue. Aim at the probability of loading. Buy guaranteed space with a penalty attached when it fails, a named vessel and a named voyage, protection against being rolled at cutoff, and a rate that does not reopen when the sailing slips. You cannot buy days in this market right now. You can buy certainty, and certainty is what the premium was supposed to be buying you.

There is a second thing nobody has said, and it carries the eight-figure price tag. The 8,007 and the 10,166 are spot numbers, and spot is where every eye is. The money is not in the spot, it is in the base those two numbers set for a January renewal. Assume a buyer moving 60 feu a month, which is 720 over twelve months, and assume the rate holds rather than collapses. The distance between 8,007 and last year's 2,290 is 5,717 dollars a box, and across 720 boxes that is roughly 4.12m dollars in a year, out of a contract that has not moved a single container differently. On the Ocean Alliance ordinary loops the arithmetic is worse: from around 4,000 before March to above 10,000 is a 6,000 dollar move, and the same 720 boxes puts that at roughly 4.32m.

Run it down to the unit, because that is where the buying decision actually gets made. Assume a 40-foot container loaded with 1,500 pieces and an ex-works price of 30 dollars a piece. A year ago the freight was 2,290 dollars a feu, which is 1.53 dollars a piece, or 5.1 per cent of the ex-works price. At 8,007 dollars it is 5.34 dollars a piece, or 17.8 per cent. The gap is 3.81 dollars a piece. To hold landed cost where it was, somebody has to find 3.81 dollars: put the retail price up by 12.7 per cent of the ex-works value, take 12.7 per cent off the factory gate price, or give up your own margin. Three doors, and the freight invoice does not care which one you walk through.

Then translate the capacity number, because 12 per cent sounds abstract right up to the moment you queue behind it. If congestion absorbs roughly 12 per cent of vessel capacity, 88 per cent is working. A buyer who needs 60 feu of space a month therefore has to book for about 68, since 60 divided by 0.88 comes out at roughly 68.2. Those eight extra boxes are not a rounding error, they are the containers you will come up short by if you book to plan instead of booking to reality. And that is the reason the roll-over, not the rate, is what will actually hurt you.

Put a price on the roll-over and the argument settles itself. Assume a SKU selling 800 units a week, a unit margin of 12 dollars, and safety stock held at two weeks, which is 1,600 units. Take a purchase order of 1,500 units, one feu, rolled three weeks. The first two weeks are covered by the safety stock. The third is not: 800 units of lost sale at 12 dollars is 9,600 dollars of gross margin gone. Against that, the freight increase on the same box is 6,000 dollars, from around 4,000 to above 10,000. One week of empty shelf costs 1.6 times the entire freight increase, and two weeks costs 3.2 times it. Now tell me which clause in your contract is worth more.

Run the counterfactual. Suppose the peak had broken on schedule in October, the way the seasonal models said it would. Carriers would have had to put capacity back into the trade rather than manage it out, spot would have fallen back towards the low thousands, and a January renewal would have been signed against a base somewhere near last year's 2,290 dollars rather than today's 8,007. Same contract, same volumes, roughly four million dollars of difference in annual freight spend. That is what an unbroken peak is worth, measured against one medium-sized importer's budget.

And the reverse branch. Suppose it runs past Chinese New Year into the spring. Then the ordinary loops stay above 10,000, the premium stays inverted, and the renewal you signed in January looks cheap by April, which is a pleasant surprise and not a plan. What is worth planning for is the shape of it. A market that prices scarcity of space instead of speed of transit behaves differently at every level, and the buyer who keeps negotiating days in a market that only sells certainty will keep winning the argument and losing the season.

Vivian Zhao looks at the same 8,007 dollars and asks whether the number will hold on the water, which is her job and she is good at it. My question as a buyer is the other half of it. Not whether it holds, but how long I have to pay it, and what I am actually getting while I do. Most buyers only get round to the second question after the contract is signed, which is why most buyers answer it with a variance report.

So if I were the one signing, here is how I would do it. Begin by separating the two products on the quote. Ask for the ordinary loop and the premium service priced side by side, with the transit times written next to them rather than left to the imagination. When the premium is cheaper and the days are the same, you have learned something worth real money, and you should stop paying the difference.

Next, move the negotiation onto certainty. Ask for guaranteed loading with a defined penalty when the box is rolled, a named vessel and a named voyage, and protection at cutoff. Put a number on the penalty rather than leaving it at best efforts: 100 dollars per feu per day of delay is easy to agree and it concentrates the mind wonderfully. Settle the free time and demurrage terms in the same sitting, because congestion is exactly where those two get quietly lost.

After that, fix the index. If the contract is index-linked, name the index, the publication, the weekday it is read on, and how often it resets. An index without a source and a date is not an index, it is a blank that somebody fills in later. If it is a fixed rate, fix it for a stated volume and say what happens above that volume, because the squeeze is precisely when you will need the eleventh container rather than the sixth.

Then deal with the calendar. Paper the January renewal in November, before the pre-holiday loading rush eats everyone's attention. Build the February arrivals into the conversation out loud instead of hoping they behave. And set a trigger you can live with: if the Drewry assessment falls below a line you choose, say 5,000 dollars a feu, the contract reopens; if it goes above 10,000, it reopens too. A contract that only ever reopens when rates fall is not a contract, it is a favour.

All told, this money comes out of my pocket, so I would rather choose what I am buying than read it on an invoice in February. If I am signing this, I pay for certainty rather than for days, I price the empty shelf before I argue about the rate, and I get the January paper finished in November, while the people who can sign it are still in the building. Winning the discount and losing the season is not a negotiation. It is a scheduling error.

↑ Back to the key points

— By Grace Yang