Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- By 20 July, split your June volume into the part that matches your January to May run rate and the part above it, and build the plan on the first number only.
- Buy the above-trend portion week by week on spot for four to eight weeks, and keep it out of any annual commitment.
- Ask your air terminal and CFS for screening and build-up slot availability in the last 48 hours before cutoff; if it is full, move your own cutoff 12 hours earlier.
- For IC and AI-chain cargo, book the bonded truck and the screening slot as one appointment; a truck holding an unscreened pallet is a parked truck.
- Track the import-export gap monthly: two more months of imports outgrowing exports means plan inland empty repositioning, not box purchases.
China's customs data showed June exports rose 27.0% year-on-year and imports jumped 36.0%, both far above forecasts, with integrated-circuit exports up 122.3%. For the first half, exports grew 17.6% and imports 26.6%. Analysts tied the strength to AI supply-chain investment and a post-oil-price-release export catch-up.
The Analysis
Read a 27% export month as a trucking diagram, not as a trade statistic. Twenty-seven percent more boxes leaving means more dock doors, more gate transactions, more screening positions and more drivers, all inside the same four weeks, with no extra month added to the calendar to absorb them.
China's June customs figures put exports up 27.0% year on year and imports up 36.0%, both well above forecasts. Integrated circuit exports rose 122.3%. For the first half, exports were up 17.6% and imports 26.6%. The stated drivers are AI supply-chain investment and a catch-up in exports after oil prices came off.
The single-category number is the one I would build a plan around, and not because it is the biggest. It tells you which part of the physical chain is about to be asked for something it does not have.
The two headline numbers are not the same shape, and the half-year numbers prove it. If exports grew 17.6% across six months and 27.0% in June, then the first five months ran at roughly 15.7%, assuming the months carry equal weight, which is my assumption and not the customs bureau's. Run the same subtraction on imports: 26.6% for the half year against 36.0% in June leaves about 24.7% for January to May. So imports are a trend with a bump, and exports are a spike on a slower trend. Capacity bought for one shape is wrong for the other.
Follow the boxes. A 27% export month starts at the factory gate, four hours before anybody thinks about a vessel: more trucks queueing for the same dock doors, more pallets into the same CFS floor, more gate transactions at the same terminal, more yard positions, more vessel slots. Every one of those is a fixed footprint in the short run. The trucking is the first to bite because it is the only step with no inventory buffer; a dock door can hold a queue, a yard can hold a stack, a vessel can hold a late box, but a truck that cannot get into the gate is a truck doing nothing at 90 dollars an hour.
The AI-chain goods change which step binds. Integrated circuits and the equipment around them are high value and low weight, so a disproportionate share moves by air, and air cargo's bottleneck is not aircraft. It is the terminal's screening and build-up area in the last 48 hours before cutoff, plus the bonded trucking window that gets a pallet from the factory to that area. A 122.3% jump in that category does not ask for more planes. It asks for more X-ray positions, more build-up floor, more security-trained handlers and more appointment slots in a window that is already the narrowest 48 hours of the week.
Layering runs along the tier line, not the size line. Tier-one AI-chain shippers with their own block space agreements, bonded warehouses and dedicated trucking ride this out; they are the ones the forwarders protect. Tier-two and tier-three suppliers shipping through consolidators get rolled, and rolled cargo misses a vessel or a flight and then waits for the next one. Consumer-goods exporters with no AI content are the quiet victims: they compete for the same trucks and the same screening slots with cargo that earns twenty times more per kilo, and they lose that competition every time. Direction matters as well. The import number, up 36.0% and running near 24.7% before June, lands on the destination side, at the marine terminals and then on the rail ramps inland; that is a steady load, and steady loads are easier to plan for than pulses.
The empty box is where the release is pointing the wrong way, and this is the part worth arguing about. Imports growing faster than exports means more full boxes land than leave, which physically adds empties back into the pool. The usual pre-peak worry about container availability is, this time, aimed at the wrong target. What replaces it is location and timing: empties return through the import gateways on the import schedule, while a growing share of export cargo loads inland and needs the box there, on the day the factory is ready. The box exists. The repositioning move that puts it in the right place does not, and that move is a truck and a chassis and a driver, which is the same constrained resource the export surge is already competing for.
Timing, in the order the phone rings. Air screening and build-up slots tighten one to two weeks after a category spike like this one. Port gate and yard congestion follows at two to four weeks. The destination rail ramps, which are already absorbing a modal shift, feel the inbound wave at three to six weeks. The batch itself clears in four to eight weeks if it is genuinely a catch-up, and that is the single most useful question you can ask about this number: is it a rate, or is it a release of deferred orders? The article's own explanation, a catch-up after oil prices came off, says release. Releases end.
Here is the arithmetic, with every assumption labelled. Suppose you ship 200 FEU a month by ocean and 40 tonnes by air. At the run rate of 15.7%, you should plan for about 31 extra FEU a month; at the June rate of 27.0%, you would plan for 54. The 23 FEU between those two numbers is the spike. If you sign a twelve-month commitment at the June level and the spike clears after two months, you are holding 23 FEU a month of paid, unfilled capacity for ten months, which is 230 FEU of commitment you cannot use. Buy the spike on spot, week by week, and let it expire with the batch.
The traps are all versions of the same mistake, which is treating a monthly print as an annual run rate. Signing a year of air capacity on a catch-up month is the expensive version. The second trap is reading the average as your cargo: 27.0% is a blend, and inside that blend one category grew 122.3% while everything else grew less than the headline, so your lane may be flat while the country is up a quarter. The third is assuming that more volume means more boxes are the constraint, when the arithmetic on imports says the opposite; you will spend money buying something that is not short and still miss the cutoff.
If oil had not come off and the deferred orders had not been released, June would have looked like the first five months, somewhere near 15.7%, and nobody would be writing about a surge. The physical system would have absorbed it without a story. The story is the batch, and the batch is the thing that breaks a dock door schedule, not the trend.
↑ Back to the key points