WorldACD: Air Spot Rates Hold at $3.55/kg in Week 41 as Jet Fuel Stays 111% Above 2025

The Analysis

A box of phone accessories out of Shenzhen is a simple thing on paper. Fly it west on a Monday night, change aircraft at a Gulf hub before dawn, and have it on a truck into Frankfurt by Wednesday. I have walked this line more times than I can count, and the part everybody argues about is the rate. The part that has actually spoiled my shipments is the transfer. So when WorldACD published week 41, I read the rate line, and then I went straight to the capacity line, because on this road the rate tells you what the market felt last week while the capacity line tells you where the queue will form next month.

Here is the week as published. Global air cargo spot rates held at three dollars and fifty-five cents a kilo over five to eleven October, unchanged on the week. Tonnage fell one percent week on week, a third weekly fall in a row, and still ran nine percent above the same week of 2025. The all-market rate, which folds contract cargo in with spot, sat at three dollars and thirteen cents, up twenty-nine percent year on year, while spot on its own is up thirty-four. Asia-Pacific origin spot slipped three percent to four eighty-three. Middle East and South Asia slipped two. Africa went the other way and rose ten. Capacity shrank one percent for a second week, and inside that number Gulf capacity fell four percent on the week and now sits twenty percent below where it was in late February. Jet fuel costs one hundred and eleven percent more than a year ago.

Flat rate, falling tonnage, falling capacity, fuel at more than double. I want to walk this the way I would walk a border crossing: which node clogs, what it costs to change the hub, change the mode or change the gateway, and whether this road is still worth driving into November and December. Every figure in that list is public. What is not in any headline is that the flat rate is the least informative thing on the page and the twenty percent missing from the Gulf is the most. If it comes down to one sentence, this is a road problem, and road problems get solved by finding the queue before it forms rather than by arguing about last week's toll.

Start with the arithmetic almost nobody bothers to run. Tonnage down one percent three weeks running compounds to roughly three percent off the top of the month. Capacity down one percent for two weeks is about two percent off the space available to buy. Two sides of the same market shrinking at close to the same pace, and the price did not budge. That is not a quiet market. That is a market being balanced by withdrawal on the supply side while demand, measured against the same week of last year, is nine percent higher. Demand is not what is falling on this road. Space is.

Why in October, and why does the calendar matter? The trimming is happening in the weeks when the fourth-quarter ramp normally begins, not in the weeks after it ends, and that is the tell. A carrier expecting a weak peak would be adding frequencies into November rather than pulling them. Put fuel into the same picture. Jet fuel at one hundred and eleven percent above last year makes the thin legs of a network uneconomic to fly, and the capacity that comes off first is always the capacity that was marginal before the fuel bill doubled. A freighter rotation that made a modest profit at last year's fuel price can be a loss-maker at this one, so it goes.

Now the layering, because nobody ships a global average. Asia-Pacific origin spot is four dollars and eighty-three cents after falling three percent, and that is thirty-six percent above the three fifty-five in the headline. If you are an exporter booking out of Shenzhen, Guangzhou, Hong Kong, Hangzhou or Ho Chi Minh, the number in the news has never been your number. Yours is four eighty-three, and the three percent dip reads like relief only until you hold it against a market where global spot is up thirty-four percent year on year and the all-market rate is up twenty-nine. A lane still running a third above last year has not gone soft because it slipped three percent in a week.

Middle East and South Asia slipped two percent, which makes it the cheapest-looking lane on the board, and it is also the lane whose transfer node has lost a fifth of its capacity since late February. Gulf capacity is down four percent on the week and twenty percent against February. Do not underestimate that port. Cheap freight out of a thinning hub is not cheap, it is simply priced before the consequence arrives.

Africa went the other way and rose ten percent, and that number deserves more attention than it will get, because a ten percent move on a thin lane is a different animal from a ten percent move on a thick one. Thin lanes have few carriers, few alternatives and long recovery times, so a rate that jumps ten percent there tends to stay jumped rather than mean-revert in a fortnight. If your boxes go into African destinations, your increase this month is ten, not three, and any budget built on a flat global rate was wrong by seven points before anybody opened the file.

Then there is the gap that matters most to whoever signs the contract. The all-market rate with contract cargo inside it is three thirteen against spot at three fifty-five, a difference of forty-two cents a kilo, roughly thirteen percent. Contracted cargo is therefore flying at a thirteen percent discount to the spot market right now, and that discount is not a bargain you have secured in perpetuity. It is the height of the cliff you walk off at renewal, and the renewal round for most importers and exporters lands in January or the moment an annual rate is reopened.

The layering also has a size dimension that the averages hide. A shipper with a proper annual contract is sitting at the three-thirteen end of the market and will barely feel October. A shipper without one, which in practice means small and mid-sized exporters booking through a forwarder week by week, lives entirely at the four-eighty-three end out of Asia-Pacific and has no protection at all against the peak. Same published data, two completely different quarters. If you are in the second group, everything below applies to you twice over.

How quickly does any of this reach a shipping desk? Capacity comes off at the edges of a network first, so the earliest sign is not price, it is a frequency that quietly disappears from the schedule and a connection that gets forty minutes tighter. On a transfer-heavy lane that shows up in transit time within one to two weeks. Price follows in two to four, because spot only moves when the space on the day you want it is genuinely gone. Contract follows last, at renewal. The practical consequence is the booking window: three to five days out was workable in September, ten to fourteen days is what mid-November will demand, and any forwarder still quoting a three-day window in the second half of November is quoting off a market that no longer exists.

Here is the part I have not seen in a single write-up of this data, and it is the part that will cost importers and exporters real money. A flat global average is not a calm market. It is an index in which rises and falls are cancelling each other out: Africa up ten, Asia-Pacific down three, Middle East and South Asia down two, all folded into a number that did not move. Nobody ships an index. You ship a lane, out of a specific airport, on a specific day, and an index being flat tells you nothing about whether your lane is the one going up or the one going down. The stability everyone is reading into three fifty-five is a composition effect, and composition effects are where budgets get quietly destroyed.

The next thing nobody printed is what the Gulf number actually does to your cargo. Capacity out of the Gulf is down four percent week on week and twenty percent against late February, and yet rates on the Middle East and South Asia lane fell two percent. Read that pairing again. It looks like a contradiction and it is not one. A transfer hub being de-peaked loses frequencies before it loses flights, and its rates soften because belly capacity once sold at a premium is now being sold to fill aircraft. What you get for the cheaper rate is a thinner connection bank.

Your cargo does not get expensive first at a thinning hub. It gets rolled. It misses the connection, it sits a day, and then it gets expensive, because by then you are buying space on the day you need it at the spot rate, which is the most expensive way anyone has ever bought air freight. I have watched that film at more than one transfer point, and it always ends the same way: the shipper who took the cheap lane pays the most.

So what does it cost to change the hub, which is the question I always ask myself on this road? On Asia to Europe quotes I have handled, moving a shipment off a Gulf transfer and onto a direct freighter has typically carried a premium in the mid-teens as a percentage of the rate, and moving it to a different transfer point such as Istanbul has usually been cheaper than that but not free. Get your own number rather than taking mine, because the premium moves with the lane and the week. What you are buying with it is a day to a day and a half of transit and, more importantly, the removal of the connection risk entirely. Against that, weigh what a roll-over actually costs you: one missed connection on a production-critical part is a line stoppage, and a line stoppage is not measured against fifteen percent of a freight rate.

Then there is the number that does not add up, and it is worth five minutes of anybody's week. Fuel is one hundred and eleven percent above last year. Spot is up thirty-four and the all-market rate is up twenty-nine. Carriers have not absorbed a doubling of their fuel bill out of goodwill, so somewhere in your invoice there is a line doing work the rate table is not doing. Whether your fuel surcharge and security surcharge sit inside the rate your forwarder quotes or on top of it is not a footnote, it is the difference between a rate increase and a cost increase. Get that answer in writing this week, from your own forwarder, on your own lanes, because every landed-cost model built on a rate that excludes the surcharge line is a model that is wrong by exactly however much that surcharge turns out to be.

Let me put numbers on it and write the assumptions down so you can argue with them. Assume you move twelve tonnes a month out of Asia-Pacific into Europe, seventy percent on an annual contract and thirty percent on spot. Take the contract portion as effectively paying the all-market average of three thirteen a kilo, and the spot portion as paying four eighty-three, because you load in Asia-Pacific and four eighty-three is the published number for that origin, not three fifty-five. Eight thousand four hundred kilos at three thirteen is twenty-six thousand two hundred and ninety-two dollars. Three thousand six hundred kilos at four eighty-three is seventeen thousand three hundred and eighty-eight.

The month comes to forty-three thousand six hundred and eighty dollars. Now let the contract roll off and put all twelve tonnes on spot at four eighty-three: fifty-seven thousand nine hundred and sixty dollars. Same cargo, same aircraft, fourteen thousand two hundred and eighty dollars more a month, close to a third on top of the freight bill. And if the budget had been built off the headline three fifty-five, it said forty-two thousand six hundred while the invoice says fifty-seven thousand nine hundred and sixty. That is fifteen thousand three hundred and sixty dollars short in a month in which the rate, by the published data, did not move at all.

Add the surcharge line on the same assumptions. If fuel and security surcharges run around thirty-five cents a kilo now against seventeen cents a year ago, and those two figures are yours to replace with your own, twelve tonnes costs four thousand two hundred dollars this month against two thousand and forty last year, roughly two thousand one hundred and sixty dollars a month of increase that never shows up in a rate index. Put the two together and a shipper budgeting off a flat headline rate is carrying something close to seventeen thousand five hundred dollars a month of unbudgeted cost on twelve tonnes, in a week when the news said rates were stable. That gap is the reason this piece exists.

Where could I be wrong, because a column that cannot be falsified is not analysis. The whole reading rests on demand holding up. If that nine percent year-on-year volume figure decays through October and November toward two percent or below, then what looks like a supply-side squeeze is really a late demand rollover, capacity stops being scarce, and spot breaks down through three twenty toward two ninety. In that world, locking sixty percent of peak volume at today's rates is overpaying and staying spot-heavy is correct. Name the tripwire before acting rather than after.

Above five percent year-on-year volume, the squeeze reading holds and you lock. Between two and five, sit at forty percent locked and keep the rest spot. Below two, stop locking and let the rate come to you. The other condition that flips the conclusion is the Gulf: if Gulf capacity recovers to within ten percent of its late February level, the roll-over risk at the transfer node collapses and there is no reason to pay a premium for direct routings or alternative hubs.

What I would actually do this week, sitting on the importer or exporter side of it. Price the lane you ship rather than the index: pull your own last eight weeks of rates and surcharges by origin airport, and if your Asia-Pacific all-in number is above five dollars a kilo, move twenty percent of November volume onto a blocked-space agreement with a fixed rate before the month is out. Space, not price, is what disappears first in a market like this, and a blocked-space agreement with a rate attached is the only instrument that buys you both at once. On the contract side, open the renewal conversation in October instead of December, because the thirteen percent gap between the all-market rate and spot is the thing that closes, and it closes on the carrier's timetable rather than yours.

The mode question deserves a proper answer rather than a reflex. For any SKU where air freight runs above fifteen percent of goods value, the air lane is not buying you speed, it is buying you a habit, and the money is better spent moving that SKU onto a sea-rail multimodal routing with a twenty-five day door-to-door that you have already tested. Test it in October, while a failed test costs you a week instead of a season. For genuinely urgent cargo the choice is not air versus sea, it is which air routing: a direct flight costing more per kilo that lands when you said it would, or a cheaper transfer that lands when the hub feels like it. In a quarter where the transfer node is being de-peaked, pay for the direct.

Inventory is the other lever, and it is the one people reach for last. A twenty-five day multimodal routing against a five to seven day air transit is an extra eighteen to twenty days of pipeline, so if you shift even a quarter of your air volume onto it you need roughly five extra days of cover on those SKUs to keep the same service level. That costs carrying charge, and on most cargo carrying charge is far cheaper than peak-season air spot. Work the number on your own inventory: if five days of cover on the shifted SKUs costs less than a third of the peak air premium you expect to pay, the stock is worth holding and the mode shift pays for itself.

The pitfalls are predictable, which is what makes them avoidable. An Asia-Pacific shipper building a budget on three fifty-five when the lane prints four eighty-three has lost before the market even moves. Another trap is treating the cheapest lane as the cheapest option, which is precisely backwards when the cheap lane is the one whose hub has lost twenty percent of its capacity since February. Then there is waiting for the spike before doing anything about space: by the time spot prints above three seventy, the space for the week you need is gone and you are buying at the top. And the one I see most often is signing a fourth-quarter contract without knowing whether the surcharge line sits inside it, which turns a negotiation about rates into a surprise about costs.

Does this road still get driven? It does, and for most cargo it remains the only road that arrives this quarter. What changes is how you drive it. Book further out than you did in September. Buy space where you need it and stop admiring the lanes that look cheap. Put your own rate, your own surcharge and your own origin into the model instead of a global average that is flat by construction. Back on the Shenzhen to Frankfurt run I opened with, the failure point was never the price of the ticket. It was the connection at the hub, and the connection is exactly where this week's data says the trouble is forming.

↑ Back to the key points

— By Ömer Kaya