USTR Moves to Freeze China Ship Port Fees Two More Months, Through 9 January 2027

The Analysis

"The notice says zero. Read the dates before you read the relief." Filed on 23 October and published in the Federal Register on 26 October, the USTR action holds the Section 301 maritime service fees at zero from 10 November 2026 through 9 January 2027, two months past the 9 November expiry of the suspension that is running today. The charges it freezes are the ones the trade has been quoting without pricing: 18 dollars per net ton or 120 dollars per discharged container on Chinese-built tonnage, 50 dollars per net ton on Chinese-owned ships, and the 100% duty on Chinese ship-to-shore cranes.

Two months is not a settlement. It is sixty days of calendar bought with four days of comment. The docket, USTR-2026-0694, closes on 30 October, four days after publication and before most of the industry has finished reading the text. A window that short tells you something about the decision that came before it: the decision was taken, and the paperwork is catching up.

So we do what this column always does. We take the number apart before the number takes the margin apart. Four figures sit on the table: 18, 120, 50, 100. Not one of them is usable in the form the trade press has been repeating it. We will check the base, then the unit, then the trigger, in that order, because that is the order in which they reach an invoice.

Begin with the document rather than the headline. What was filed on 23 October and printed on 26 October is a suspension of a suspension. The Section 301 maritime action carries the Annex I, II and III service fees, and those fees carry two rates: 18 dollars per net ton or 120 dollars per discharged container on a hull built in China, and 50 dollars per net ton on a vessel a Chinese operator owns, wherever the steel was welded. The 100% duty on Chinese ship-to-shore cranes rides in the same freeze. What expires on 9 November is not the charge. It is the zero. What the new notice does is move that zero to 9 January 2027.

Write the four numbers down in the order in which they will come looking for you. Eighteen: dollars per net ton, on build. One hundred and twenty: dollars per box discharged, on build, and this is the figure that bites container cargo. Fifty: dollars per net ton, on ownership. One hundred: per cent, on a crane. Those are four different animals, and the summaries have been running them together as though they were one charge with four spellings. That habit is the reason nobody has priced this properly.

Why freeze again, and why sixty days? Two reasons, and both are real. One is calendar. The 9 November date sits close to the wider tariff truce timetable, and a lapse would have detonated a shipping charge in the middle of a negotiating sequence where neither capital wants the freight invoice to be the thing that tips the table. Sixty days pushes the decision past the holiday booking season and into a week when the people who can sign are back at their desks. The other reason is domestic arithmetic, and it is the more interesting one. A fee that bites at the American end of the trade bites American operators, American ports and American shippers in the same month, and all three of them keep offices in Washington with people who know how to use them.

Port authorities, agricultural exporters and importers' associations spent the summer saying the same thing in different rooms. A fee on Chinese-built tonnage is a fee on most of the fleet, because Chinese yards have touched most of the fleet. A fee on ship-to-shore cranes is a fee on American port capital budgets, because no alternative supplier delivers that tonnage of steel and that control software next year at any price. When the parties on the receiving end are your own ports and your own growers, the charge stops behaving like a trade measure and starts behaving like a domestic cost line. Guess whose invoice it lands on. The answer is not the operator, whatever the operator tells you on the phone.

Now the order of pain, because it is not uniform. Carriers collect. They do not pay. The charge is assessed against the operator at the moment of entry and reappears on your bill three to six weeks later under a name your accounts payable team will not recognise. Chinese-owned operators carry the heaviest load, because 50 dollars per net ton on ownership sits in a different order of magnitude from 18 dollars per net ton on build. Chinese-built tonnage under foreign ownership carries the lighter one. Hulls built in Korea or Japan and owned outside China carry nothing at all, which is precisely why the strings get reshuffled before the first dollar is ever assessed. Watch the vessel lists rather than the press releases: the redeployment shows up on the rotations weeks before it shows up in any filing.

Importers of boxed cargo meet this as a per-box charge, and that is why 120 outranks 18 in your budgeting. Exporters of bulk and agricultural cargo meet it sideways, through space and equipment. When a carrier pulls a Chinese-owned ship off a United States string to dodge the fee, capacity leaves the trade you move on, and your rate rises on cargo that has nothing whatever to do with the fee. Forwarders and non-vessel-operating common carriers sit in the middle and finance the gap, which matters more than it sounds: a fee assessed at entry falls due weeks before the matching surcharge can be collected from you, somebody has to fund that interval, and the cost of funding it is already inside the quote you are holding. The one party that feels nothing is whoever is moving cargo today under a fixed annual rate that has not yet expired. For that shipper this is a January problem, which in October is the best possible kind of problem to have.

Three dates, and only one of them should change what you do today. Start with 30 October, the close of comments on USTR-2026-0694. Miss it and your view is not in the record, though on a four-day window you should not build a plan around having a say. Then 10 November, the day the new zero opens and the day the old suspension would have died. The same day, which is the tidiest thing about the drafting. Then 9 January 2027, and this is the date that belongs in a contract clause rather than in a diary. Use the docket while it is open, because there is one question worth putting on the record: who elects between the net ton basis and the container basis, when is that election made, does it bind for a period or voyage by voyage, and over what unit is the resulting charge assessed. Four parts, one sentence, and the answer decides whether a single West Coast rotation costs 300,000 dollars or 900,000.

A charge of this shape is assessed at entry, not at booking. The exposure on a shipment is therefore set by the day the vessel arrives, not the day you signed the booking note. Any agreement you sign in November covering cargo that lands in the first quarter of 2027 has to state who carries a charge that reappears on 10 January. Leave it silent and the answer will be supplied later, by whoever is holding the invoice.

Cranes run on a slower clock, which is exactly why they get ignored. A 100% duty on a Chinese ship-to-shore crane does not touch a freight bill next month. It lands on a terminal's capital budget the moment an order is placed, and it returns as terminal handling cost two or three years later, spread across every box that terminal lifts. That is the slowest number in this notice and the one nobody will ever trace back to it. If you lift at a terminal with Chinese cranes on order, ask what the duty does to its capital plan and when that reaches the tariff. The answer will not be pleasant, but eighteen months of warning beats eighteen days of it.

Here is what has been left out of the summaries, and it is not a footnote. Every report repeats the two rates on Chinese-built tonnage as though they were a menu: 18 dollars per net ton or 120 dollars per discharged container, take your pick. Nobody has said who picks, what the pick is between, or over what unit the charge is counted. Those three questions are worth more to your budget than either rate. They are not academic either. Answer them and the gap on one vessel across one year of West Coast rotations runs to seven figures, which is roughly the size of the line your carrier will write into the next renewal. Leave them unanswered and you are not pricing a fee, you are guessing at one.

Let us run the arithmetic on a plain mid-size box ship. Assume a nominal 8,000 TEU vessel, assume net tonnage of 45,000, and assume it discharges 2,500 boxes at a single United States port call. Those are our assumptions: substitute your own hull and argue with the result. Per net ton, 45,000 times 18 is 810,000 dollars. Per container, 2,500 times 120 is 300,000 dollars. A rational operator elects the container count and keeps 510,000 dollars on that one call. The break-even sits where 18 times net tonnage equals 120 times boxes, which is 6.67 net tons per discharged box. Our assumed hull runs at 18 net tons per box, far above the line, so the per-box election wins. So does nearly every container ship in the trade, because a box boat carries far more than 6.67 net tons for each container it actually discharges. The 18-dollar headline is decoration for container cargo. The figure that binds is 120 dollars a box.

Then the unit. Is that 120 dollars charged once per voyage, once per port entry, or once per vessel per year? Three readings, three very different invoices. Take the same assumed ship on a West Coast rotation calling three United States ports: 2,500 boxes times 3 calls times 120 dollars is 900,000 dollars per voyage, and twelve voyages a year puts it at 10.8 million. Assess it once per vessel per year instead and the same hull owes a fraction of that. The notice as reported does not settle which reading applies, and the distance between the two is close to an order of magnitude on an operator's annual bill. That distance is the difference between a surcharge you absorb and a surcharge you fight.

And then the asymmetry between building and owning, which is the part that should change how you pick a counterparty. Put the same assumed 45,000 net ton hull under the ownership charge rather than the build charge: 45,000 times 50 is 2.25 million, against the 300,000 dollars the build election would have produced on the same call. The ownership fee is seven and a half times the build fee on an identical ship. That ratio is the entire policy in one line.

Whatever the stated purpose, the structure penalises who owns the vessel far more heavily than where it was built, which tells you precisely who will reshuffle fleets and who will simply pay and then bill you for it. There is a mechanical joint here that decides who really pays, and it sits in your contract rather than in the regulation. A charge assessed on the operator and a charge borne by the operator are not the same thing; the surcharge clause is what sits between them.

If your agreement lets the carrier pass through any charge imposed by any authority, the fee arrives on your invoice automatically and there is nothing left to discuss afterwards. If it names what can be passed through and puts a ceiling on it, you have a negotiation. Most paper in this trade was written the first way, which was a free drafting choice at the time and is not free now.

Ingrid Voss, who reads texts like this for a living and dislikes being told what they say, would put the same point more carefully: the applicable condition here is the definition, and the definition is carrying more weight than the rate. She is right, and it is worth being blunt about the reason. Built in China is a fact about a hull, and it never changes. Owned in China is a fact about a corporate structure, and it changes with a share transfer, a bareboat charter or a flag. One of those two charges can be re-engineered by lawyers. The other can only be paid, or avoided by taking the ship off the trade entirely. That asymmetry decides who ends up carrying the fee long before the rate gets a say.

Run it backwards, because the counterfactual is what tells you what the freeze was worth. Had nothing been filed, the zero would have expired on 9 November and the charges would have gone live into the middle of the peak. Operators would have had two options. Pay and pass through: on our assumed hull that is 120 dollars per discharged box turning into a named surcharge inside a quarter, 900,000 dollars a voyage across three port calls, and a procurement team explaining a line item nobody budgeted for. Or redeploy: Chinese-owned tonnage leaves United States strings, capacity comes out of a trade that was already tight, and the spot rate rises by more than the fee was ever worth. Either way the shipper pays. Only the wording on the invoice changes, and the wording is the part you could have argued about in November.

The forward branch matters more, because both of its outcomes are still live and one of them gets decided before the year ends. If the freeze is renewed in January and renewed again after that, the fee stops being leverage and becomes a memory, the crane duty stops hanging over capital budgets, and the fleet reshuffling already carried out in anticipation turns into a stranded cost that sits inside your rate for years. If it lapses on 10 January 2027 instead, then every contract signed this winter without a fee clause becomes a January invoice with your name on it. You do not get to pick which branch happens. You only get to pick which one you have priced, and pricing both of them costs you an afternoon with your own twelve-month volume and a calculator.

So what goes into the paper, and by when. Begin with the fee clause in your next agreement and read it on the assumption that the charge goes live on 10 January 2027. That assumption costs you nothing if you are wrong and protects your margin if you are right, which is the cheapest insurance available in this trade.

Next, count your port calls rather than your bookings. Take every United States arrival you moved in 2026, list the entry dates and the boxes discharged at each, then run them at 120 dollars a box for the build exposure and 50 dollars a net ton for the ownership exposure. Do it by string rather than by customer, because the fee attaches to the vessel, and the vessel does not know whose cargo is inside it.

After that, put two questions to your carrier in writing and keep the answer. Which election do you intend to make on a Chinese-built hull? Over what unit is the charge assessed: per entry, per voyage, or per vessel per year? A carrier that will not answer those two questions in writing has already told you where the fee is going to land.

Then price both branches before you sign anything. Take the January lapse case and put a number on your first-quarter exposure. Take the rolling-renewal case and ask what you would actually be paying for in a rate that never resets. On the assumed 8,000 TEU hull, the build exposure alone runs to 10.8 million dollars a year under the per-entry reading, across three port calls and twelve voyages, and that arithmetic is what your carrier's surcharge will be built on whether or not it is ever assessed.

And put a date on the renewal conversation. Get the January contract onto the table in the first half of December, while the people who can sign it are still answering their phones. Leave it later and you are negotiating with an out-of-office reply, and the out-of-office reply has never once conceded a point.

Circle 9 January 2027 and treat it as a hard date, because this trade has a long habit of turning a tariff into a surcharge and launching it during your busiest week. Two months is not a resolution. It is a calendar entry that hired better lawyers. If the rate needs locking, lock it before the middle of December: the number you agree then is the number you pay for the whole spring.

↑ Back to the key points

— By Vivian Zhao