Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- By 30 September, ask every forwarder you use which Q4 allocations on your two main lanes were released and by whom, and put a name and a tonnage next to each one.
- Split Q4 tonnage into a protected slice with a named allocation and an unprotected slice, and cap the unprotected slice at 20% of the total so one late tender cannot take the promise down with it.
- For time-critical freight, write the cutoff hour into the contract instead of the transit time, and review that clause monthly from October through January.
- Set a minimum of five pallets or one ULD per booking on the two lanes where fuel bites hardest, and route anything smaller into a consolidated weekly build.
- Open the 2027 allocation conversation in October rather than January, and anchor it on tonnage and cutoff hour rather than on the yield number.
IATA trimmed its 2026 air cargo volume forecast to 71.7 million tonnes, up just 0.2% year on year versus a 2.6% call in March, citing Middle East disruption. Cargo revenue is still projected at USD 162 billion, up 7.2% on 2025, with yields rising 6.5% and fuel costs jumping nearly 40% to USD 350 billion. Tight capacity keeps rates elevated for time-critical and e-commerce shipments.
The Analysis
IATA has cut its 2026 air cargo volume forecast to 71.7 million tonnes, a 0.2% increase on the year. In March the same body was calling 2.6%. The stated reason is disruption in the Middle East. Revenue is still forecast at 162 billion dollars, up 7.2%, with yields up 6.5% and the fuel bill up nearly 40% to 350 billion.
Put those two sentences side by side and you get a physical picture rather than a financial one. Almost no extra kilos are being produced this year, and considerably more money is chasing each one.
For anyone booking air space on a repeating schedule, that is not a demand forecast. It is a statement about who gets displaced, and in what order.
Start with the arithmetic nobody printed. Two tenths of a percent of 71.7 million tonnes is about 143,000 tonnes, for the entire world, for the entire year. Spread across 365 days, the net growth of air cargo volume on this planet is roughly 390 tonnes a day. That is the size of the additional pie. Everything else that anybody gains this year, somebody else gives up.
Which explains why the volume number barely moves while the money number moves a lot. A tonne of air cargo is not a uniform thing. The market added almost no tonnes and 7.2% more revenue, and that combination only works if the mix shifted toward freight that pays more per kilo and the price of the scarce slice rose. Yields up 6.5% on flat volume is arithmetic, not sentiment. The scarce slice is the one with a clock attached: time-critical industrial parts, and e-commerce parcels carrying a delivery promise.
Now the part I actually work with. Aircraft are not the binding constraint. Ground handling is. Build-up space at the origin terminal, ULD and pallet stock, screening lanes, truck slots at the cargo gate, and the cutoff hour at the handler. Those things scale with headcount and floor area, and neither gets added because a trade association revised a forecast by a tenth of a point. Flat volume means nobody is building a terminal for you. The capacity you get in Q4 is capacity that somebody else releases.
The fuel line deserves a second look, because it is written up as a cost and it behaves as an instruction. A freighter burns close to the same fuel at 60% load as at 95%. When the fuel bill rises nearly 40% to 350 billion dollars, the operator's rational move is not to fly more cheaply. It is to fly full, or not fly. That is consolidation behaviour, and it has a specific victim: the shipper with three pallets. Any carrier will prefer one customer with thirty pallets to ten customers with three, because the handling, the paperwork and the cutoff all cost the same per shipment regardless of weight. In a flat-volume, high-fuel year, the small tender loses the argument before it ever reaches the rate desk.
Who feels it first is not who you would guess. A shipper with a block space agreement and a named allocation feels almost nothing until the contract review, and then feels all of it in one afternoon. A shipper tendering spot, especially inside the last forty-eight hours before cutoff, is the one who gets bumped, and the bump does not look like a refusal. It looks like a later flight number and a slightly apologetic email.
Size works differently here than in ocean freight. Small shippers are squeezed less on price than on consistency: they get space most weeks and lose it in the two weeks that matter. Large shippers are squeezed at renewal, where a 6.5% yield number gets applied to everything they move, including the cargo that was never scarce in the first place.
Trade direction adds another turn. Exporters out of Asia into North America and Europe sit on the lanes where e-commerce has already eaten the belly space, and that is where the cutoff is real rather than theoretical. Importers moving the other way, or moving heavy industrial goods, compete for main-deck freighter space that has not grown at all, and their alternative is ocean, which is a four-to-six-week decision rather than a Tuesday decision.
When it lands matters as much as what it says. A forecast revision is a planning document, not an operating one. The Q4 space you are about to use was committed months ago on annual and seasonal allocations, so this number will not change this peak. What it changes is next year's allocation round, and the one after. If you plan 2027 capacity on it, you bid early. If you read it as a comment on this December, you read it wrong.
Here is the thing the coverage will not tell you. In a market growing at 0.2%, your growth has to be bought as displacement, which means the negotiation is not really about price. It is about who on your lane is leaving. Flight and allocation decisions are made against contracts, and a contract only frees up when somebody declines to renew. So the useful question this month is not what the rate will be in December. It is which competitor, or which neighbouring shipper on the same lane, is reducing volume into Q4, because that release is the only new capacity that will exist.
The second unspoken point follows from the fuel mechanics. Nearly every write-up of a fuel number treats it as a cost to be passed through. At the operator level it is a load-factor instruction. High fuel tells the airline to consolidate, and consolidation removes the small shipper from the flight before it removes the large one. If you tender small, the fuel number hits you twice, once in the price and once in the space.
A small account, all assumptions mine. Say you move 600 tonnes a year and you need 15% more than last year in Q4, which is 90 tonnes. Against a world net addition of 143,000 tonnes, your 90 tonnes looks trivial, and globally it is. On your lane it is not, because the world's growth is not a pool you can draw from. Now say a fifth of that 90 tonnes is the time-critical slice that cannot slip, 18 tonnes. Tendered late, those 18 tonnes take the later flight, and every one of them costs you a day. Eighteen tonnes losing a day is not a rate problem, it is eighteen customer conversations. I am not going to put a price on it, because the release gives no rate and I will not invent one. The unit here is tonnes and days, and that is enough to make the point.
If you do nothing, the market does the work for you, in the least convenient order. Your spot tenders get later flight numbers first, then your block space gets re-cut at renewal, then the expedite budget quietly absorbs the difference. None of those three events produces an alert. The counterfactual is cheap: one email in September asking a forwarder which allocations on your lane were released for Q4, and one contract clause that names the cutoff hour instead of the transit time.
The trap is treating a forecast revision as a signal about price. It is a signal about availability. Price in this market is set by the scarce slice, and the scarce slice is set by the clock, not by the average. Watch the cutoff hour at your handler. The week it moves earlier is the week the average stopped mattering.
↑ Back to the key points