Xuzhou rail-sea intermodal surges 36% in first half

The Analysis

Xuzhou is not a port. That is the point of the story, and it is the part that gets lost in the headline. The city sits inland in Jiangsu, so every box that leaves it for the sea has to be railed to a coastal terminal before it can ever touch a ship. In the first half of 2026 that channel moved 13,642 TEU, up 36.47% year on year, and 4,300 TEU of it went out under the CCA single-document model, up 249.03%. A government-enterprise task force working door to station cut one light-industry shipper's container cost by nearly 55% by rerouting that firm onto a China-Kyrgyzstan-Uzbekistan rail-road service.

Three numbers, three different kinds of number, and most of the coverage treats them as if they were the same kind. One is a volume. One is a share of a small base growing fast. One is a single company's single cost line on a single routing. If you are an exporter in northern Jiangsu or Anhui trying to decide whether next season's boxes go down this channel, the difference between those three matters more than any of the percentages.

Lay the physical chain out before arguing about the growth rate. A container leaves a factory door on a truck, enters a rail yard, gets lifted onto a wagon, rides several hundred kilometres to a coastal terminal, gets lifted off, waits in the yard for a vessel slot, and gets lifted again onto the ship. Five lifts, four waits. Intermodal deletes none of them. It moves the waiting inland, where ground is cheaper and the queue shorter, and swaps a long truck haul for a rail haul that costs less per box-kilometre. That is the whole trick, and there is nothing digital in it.

The constraint on a channel like this is almost never the track. Jiangsu has track. The constraint sits at the coastal handoff: whether the train lands inside the vessel's cut-off, whether the terminal has a yard slot when it does, and whether there was an empty box at the inland yard to load in the first place. That last one gets forgotten. Export boxes and import boxes are not the same population of steel, and an inland channel whose export leg outgrows its import leg has to drag empties back inland against the flow. Somebody pays for that drag. It hides in the container cost line, and container cost is not total cost.

The causal chain in the Xuzhou figures runs through paperwork and geography at once, which is why it reads as one story and is really two. The single-document model covers a journey crossing rail, road and sea with one contract, and it grew 249.03% to 4,300 TEU. Growth like that off a base of a few thousand boxes is what happens when a process step stops holding the cargo. The 36.47% on the total is the steadier number: 13,642 TEU across twenty-six weeks is roughly 525 TEU a week, and assuming a block train of about 50 FEU with two TEU to the FEU, that is about five trains a week. My assumption, not a published figure. Five trains a week is a real channel. It cannot absorb a whole province diverting in one quarter.

The leg toward Kyrgyzstan and Uzbekistan has different physics. Rail-road means the box gets transloaded somewhere, and transloading means a crane, a yard, a customs form and a day. It also means a mountain road section with a seasonal ceiling a railway does not have; winter closes options on that kind of route in a way it never closes a rail line. The 55% is real for the firm that got it, but it is a saving on one routing under one negotiated arrangement, brokered by a task force that cannot work door to door for every shipper in the province.

A light-industry exporter shipping full containers on a stable weekly rhythm, with a factory inside the catchment of the inland yard, gets a real cost cut and a more predictable gate. A small shipper moving two or three boxes a month gets the headline and none of the terms, because a door-to-station service is staffed for accounts with volume. Importers get less than exporters: inbound boxes get railed inland full or empty, so an import-heavy shipper on this corridor is effectively paying for the export boom's empty repositioning. Direction explains more of the outcome than company size.

The half-year figures are a rear-view mirror published in July about a period that ended in June. The forward effect lands between August and October, when vessel slots and empty boxes tighten at the same moment. That is the month an inland channel's reliability gets tested, because a shortage of either turns a five-lift chain into a six-day wait. On a thirty-day peak plan, the earliest sailings are already decided and the rest are still open.

Here is what the coverage did not say. The 249.03% is the quoted number, but 4,300 against 13,642 means the single-document model is still under a third of the channel, about 31.5%. The fast-growing number is the minority product moving through a majority legacy process. The practical consequence is that your boxes sit in the same physical queue as everyone else's, but your paperwork runs on a different clock. The failure mode is not a lost box. It is a box that arrives and cannot be released because the release document is still on the old path, eating storage while the market picks up. The other omission is that 55% is measured against container cost, not landed cost. If container charges are 18% of your all-in spend on that lane, a 55% cut is about 10% off the total. Both percentages are my assumptions for the arithmetic, not numbers from the report.

Run it plainly. A shipper spending 100 units of all-in cost per box, 18 of them container-related, lands at roughly 90 after the 55% cut. Real gain, worth chasing, and it does not change which market you can sell into. Ten points of landed cost moves margin, not a sourcing decision, and anyone reading 55% as a total-cost number will over-plan the season.

Without the task force that firm is still trucking to the coast, and the difference does not vanish, it sits in the trucking line where it always sat. Without the single document those 4,300 TEU still move, just slower, and the headline prints in the twenties instead of at 36%. The counterfactual is not cargo staying home. It is cargo going by a more expensive, less predictable road, with the province paying in dwell time rather than in rates.

The trap is treating a government-brokered pilot as a published tariff. Rates and service levels are negotiated per account, they expire, and they rest on a rail slot that is allocated rather than published. Get the allocation in writing and ask what happens to your boxes in the week the empties run short. That question, asked in July, is worth more than every percentage above. Today's action is dull: call the yard and ask how many empty forty-foot boxes are on the ground this morning.

↑ Back to the key points

— By Ömer Kaya