Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- By 15 July, pull every July and August booking confirmation and flag the ones riding a sailing that has not been blanked yet; if more than 20 percent of your third-quarter volume sits on unconfirmed sailings, rebook half of it onto confirmed space before the month ends.
- Set a dwell trigger at 4.5 days for Long Beach and Los Angeles combined: the day the published average crosses it, shift your next two weeks of discretionary drayage to rail and tell your customers in writing within 48 hours.
- For the next trans-Pacific contract round, ask for the empty-repositioning cost as a separate line rather than folded into an all-in rate; a carrier that refuses to break it out is telling you how much of it you are already paying.
- If you export westbound out of San Pedro Bay, book equipment against the import arrival schedule rather than your own production schedule; the boxes are there, the sailings are the scarce item.
- Track the port's monthly stability-window reading alongside your own on-time metric through September, and treat a widening gap between the two as an inland problem rather than a terminal problem, because at 3.2 days it almost certainly is.
The Port of Long Beach moved 932,179 TEU in May 2026, its fourth-busiest May on record, with loaded imports up 6.4% year-over-year to 466,800 TEU and exports rising 3.1% to 122,500 TEU. dwell times averaged 3.2 days and the terminal-queuing metric stayed inside the “window of stability” declared by the port earlier this month. June volumes are tracking toward 900K TEU, supported by a wave of pre-July-1 frontloading, although carriers warn that the pull-forward could flatten July/August. The Port of Los Angeles reported 880K TEU for the same month, putting the San Pedro Bay complex at 1.81M TEU combined.
The Analysis
Lay the Pacific flat and the busiest line on it this month is not a route. It is a date. Nearly everything that crossed the San Pedro Bay wharves in May and June was, in some sense, running toward the first of July.
The Port of Long Beach moved 932,179 TEU in May 2026, its fourth-busiest May on record. Loaded imports rose 6.4 percent year over year to 466,800 TEU, loaded exports rose 3.1 percent to 122,500 TEU, average dwell came in at 3.2 days, and the terminal-queue measure stayed inside the window of stability the port had declared earlier in the month. June is tracking toward 900,000 TEU. Los Angeles reported 880,000 TEU for the same month, which puts the two ports at 1.81 million TEU combined.
As a volume story that is a good month. As a network story it is a system that has learned to breathe in pulses, being asked to breathe in pulses it does not set. The useful question is not how many boxes moved. It is what the shape of that number tells you about where the next constraint surfaces, and when it reaches your booking.
Take the number apart before you trust it. Of the 932,179 TEU that crossed those wharves, 466,800 were loaded imports and 122,500 were loaded exports. What remains, roughly 342,900 TEU, a little over a third of everything the port handled, is empty equipment rather than cargo. That figure is arithmetic on the port's own reported lines, not a separate disclosure, and it is the most important number in the release. A gateway where 37 percent of the box count is empty steel is not a two-way exchange. It is an intake valve with a return pipe attached, and the return pipe runs light.
Divide the two loaded figures and you get about 3.8 import boxes for every export box. Almost everything downstream follows from that ratio. Equipment at the export gate will feel plentiful, because the empties must go somewhere and the cheapest place to put them is under a westbound booking. Sailing frequency is a different matter, and it is set by the headhaul, which means the export side inherits the schedule the import side pays for. An exporter reading a 3.1 percent export gain as a recovery should read the 6.4 percent import gain first. The gap decides how often the ship calls, and it decides how much of the empty return you are quietly funding at contract renewal.
There is a detail worth more attention than the headline. The port moved near-record volume and average dwell still sat at 3.2 days, with the queue inside its declared band. That combination means the terminal is no longer the binding constraint. When dwell stays flat while volume climbs, the buffer has moved somewhere else: inland, into rail lifts and chassis and drayage appointments, and into the empty yard, where 342,900 TEU of nothing has to be sorted, stacked and evacuated fast enough not to strangle the working stacks. The port is behaving like a corridor now, and a corridor is only as good as the road past the gate.
Who feels this first depends on which side of the contract you sit. A large importer with dedicated rail and a named account team is largely insulated, since 3.2-day dwell is simply the number its planner has assumed for two years. A mid-size importer living on allocated space feels none of this in May and all of it in July, when the sailings that carried the frontload get blanked and its bookings begin to roll. Exporters get the equipment ease without the frequency. The two directions are not symmetric either. Eastbound trans-Pacific is the headhaul and absorbs the schedule; the westbound leg and the Latin American connections inherit whatever capacity remains after the headhaul is served.
The timing reads clearly if you treat it as a wave rather than a level. May is already spent. Had May been the problem it would have shown up as a queue, and no queue appeared. June at 900,000 TEU is a 3.5 percent step down from May, which on its own is noise, but June is the last month before the date that caused all of this. The felt change arrives in the first fortnight of July as blank sailings and rolled bookings, then again in late August, when inventory that was pulled forward fails to be reordered because it is already sitting in a warehouse in Ohio. The August hole catches people out precisely because it looks like soft demand when it is borrowed demand coming due.
What the coverage will not tell you is that the stability window has become a product. The port did not merely report a queue metric; it declared a band and then reported that it was inside it. That is a different kind of disclosure from a throughput figure, and it changes what a port is selling. A landlord sells space. A port publishing a stability band is selling predictability, which happens to be the only thing an importer on a five-week replenishment cycle can contract against. Watch whether other gateways copy the format. If they do, the next decade of port competition gets fought over published service bands rather than berth depth, and the winners will be the ones whose inland leg can hold the band once the boxes come back.
Run a small calculation on your own book, with my assumptions stated plainly because the port has not published them. Take a mid-size importer moving 200 FEU a month eastbound into San Pedro Bay. Assume, and this is my assumption rather than a published figure, that the July 1 date pulled 8 percent of third-quarter volume into June. That is 16 FEU a month of demand that will not be there in August. It sounds trivial until you lay it against fixed commitments: your minimum quantity does not fall by 8 percent, your warehouse lease does not fall by 8 percent, your drayage contract does not fall by 8 percent. The volume was borrowed. The commitments were not. That gap is what shows up as cost, not the box count.
The trap in this release is reading 932,179 as demand. It is partly empties, partly a date, and partly a port reporting a gross figure when the useful figure is net loaded boxes. The second trap is treating June's 900,000 as the peak. It is the crest of a wave manufactured by a calendar, and calendar waves have a trough on the far side that runs deeper than the crest ran high, because the inventory has to be consumed before anyone orders again.
Ten years ago this pair of ports sold position: the shortest land bridge from Asia to the American interior, and the deepest water on the coast. Position still matters, but it is no longer the product. What San Pedro Bay sold in May was rhythm, a machine that took near-record volume and held dwell at 3.2 days because its inland leg and its empty yard were tuned to absorb the pulse. The map is not changing shape this month. The clock the map runs on is.
↑ Back to the key points