Ben Gurion Airport Cargo Down 12.75% Year-to-Date as Market Stabilises

The Analysis

I start almost every one of these from the buyer's desk, so let me be honest about my first reaction when I saw the Ben Gurion figure. The headline read that air cargo at Israel's main gateway was down 12.75% for the year to date, and half my feed translated that into air freight finally cooling off. I do not buy it. What the number actually tells me is that a slab of capacity I might depend on has already moved to somewhere I can see, and before I commit to anything for the second half I want to know whether it is coming back.

The raw figures first. Ben Gurion moved 30,888 tonnes of cargo in May, against 31,391 tonnes in May 2025, down 1.6%. For the five months to May the airport handled 134,677 tonnes versus 154,372 tonnes a year earlier, which is where the 12.75% comes from. Now look at the shape. January ran up 11.1%, February up 15.0%, March fell hard, April recovered, and May landed roughly level with last year. So the entire year-to-date shortfall was dug in a single quarter, and the past two months have been close to flat.

That shape is the whole story for a buyer, and it is the opposite of a slow bleed. If demand had genuinely rotted, I would lean back, run stocks down and wait for space to get cheap. If instead this is a disruption that has mostly already happened, then the space is not cheap, it is displaced. And the window to lock my lane closes as the market-is-stabilising story gets around.

Here is why I do not trust the cooling read. Underneath the May total, dedicated freighters carried 22,892 tonnes, close to 74% of everything the airport handled that month. Passenger aircraft carried only 7,996 tonnes. This is a freighter-heavy gateway, and that one fact rewrites the risk for anyone buying space out of it. Belly capacity rides on passenger schedules, so it is relatively sticky; it does not vanish overnight. Freighter capacity is contractual and movable. An operator can lift a rotation off one corridor and drop it onto another within a matter of weeks if the rate or the risk tells it to. So a quieter node is not a cheaper node. It is a node whose capacity has been relocated, and the buyer who was counting on it is usually the last to hear.

The split inside the passenger numbers is even more telling from my side of the table. Inbound cargo on passenger aircraft rose 10.1% year on year to 4,373 tonnes, while outbound passenger cargo fell 11.1% to 3,623 tonnes. More coming in, less going out. If I source from that region and ship product out through it, my export-side belly capacity is thinning while the inbound side fills up. That is a supply asymmetry I have to price into replenishment, because the return leg of a move is getting harder to find, and legs that are hard to fill are legs that get priced accordingly.

Zoom out and the picture is not air freight softening either. Global air cargo demand is still running high; what is changing is that the map is diverging. Shippers are steering around contested airspace and capacity is chasing them. For a buyer, the old assumption that a volume dip somewhere means cheaper rates for me no longer holds. Corridor concentration is now a line in my cost model, not a footnote. Two buyers can be paying very different landed costs for the same product in the same week purely because they sit on different corridors.

Why did the hole open when it did? The sequence tells you. January ran up 11.1% and February up 15.0%, and then March fell hard. That is the signature of an outside shock, not a demand cycle. Regional conflict disruption forced operators to pull or reroute capacity, and the volume that had been flowing through the node simply stopped. By April and May the network had largely re-plumbed itself elsewhere, which is why the monthly figure came back to roughly flat. The stabilisation in May is real, but read it for what it is: the redeployment has finished for now. It is arithmetic settling, not demand healing.

Who feels this, and by how much? Take the big buyers on dedicated freighter contracts first. They are the most protected on price, because their capacity sits inside an agreement, but they are also the most exposed to a quiet reassignment, because when an operator optimises its network the contracted rotation is one of the easiest things to move. Mid-sized buyers who lean on belly space and spot get squeezed next: less belly on the outbound, fewer places to hide, and a spot market that reprices the moment the corridor looks tight. Small sellers moving high-value, low-density, time-critical goods carry the sharpest pain, because they cannot switch modes without gutting their margin.

The category angle matters too. If what you move is heavy and slow, none of this touches you; you were never on that air corridor. If what you move is light, expensive and clock-driven, consumer electronics on short launch cycles, fast-fashion replenishment, service parts sitting behind a delivery promise, then this node's behaviour is close to your whole cost structure. The buyer who feels it least is the one whose product can wait six weeks at sea. The buyer who feels it most is the one whose customer notices a two-day slip.

If I am honest about how this lands internally, it is not a sourcing problem first, it is a scorecard problem. My suppliers are scored on price, on-time, quality and, increasingly, resilience, and most scorecards still treat delivery geography as a constant. It is not. I would add a corridor-concentration line to the scorecard this quarter: what share of a supplier's committed volume rides on any single gateway, and how fast could that gateway lose capacity before the supplier even tells me. The suppliers who can answer that well are the ones I want carrying my risk. The ones who cannot are a quiet liability I am currently holding for free.

How long before this shows up in my numbers? The shock already happened in the first quarter; the repricing is what is still moving. The current stabilising framing is exactly what anyone selling space will lean on in the third-quarter negotiation season, because a market that has stopped falling is a market where a seller can argue the floor is in. That gives me a clean timing read: the pressure on my next contract conversation lands in the next few weeks, not at year end. If I wait for the picture to get clearer, I will be negotiating after the narrative has already set the price.

Now the part I think almost nobody is looking at. The number that should worry a buyer is not the 12.75% at all. It is the 74% freighter share. When a gateway owes three-quarters of its cargo to dedicated freighters, its capacity is mobile by design; it can be gone in weeks and back in weeks. So the real exposure is not my node got quieter, so I must be getting a deal. It is my node's capacity can leave before I notice, and nobody calls me first. The stabilisation everyone is celebrating is, read plainly, the sound of the last rotation finishing its move. That is the line I would circle, and it is not in the headline.

There is a second thing the headline hides, and it is about leverage rather than price. A node that has shrunk 12.75% year to date is a smaller pool. If my own volumes held flat while the pool shrank, then relative to the operators running that lane I am suddenly a bigger fish in a smaller pond, which, oddly, can improve my talking position at that specific node even as my options elsewhere shrink. But the reverse is just as true: the moment my volumes fall in line with the market, that leverage evaporates. So whether a shrinking node helps or hurts me depends entirely on whether my own flow held. I check that before I open my mouth in a negotiation.

What would flip my read? If the conflict re-escalates, May's flat print becomes a pause, not a floor, and the second-half assumption of roughly normal is wrong. If instead regional tension keeps fading, capacity may drift back into the node and the outbound belly gap closes, in which case waiting to lock space is defensible. The condition that settles it is not the monthly cargo figure; it is whether the corridor stays open long enough for operators to make it their planning base again. That is a risk call, and it is the one thing I cannot hedge with a signature.

So let me do what I actually do: take this back to the cost structure. A landed cost has the goods, the freight, the insurance, the duty, and the carrying cost of time. Air capacity touches only two of those directly, freight and the cost of time, but it contaminates all of them through the buffer I have to hold. If my replenishment plan assumed a certain air capacity at a certain rhythm, and the export side of that corridor just lost a tenth of its belly space, then my plan is quietly under-buffered. And an under-buffered plan does not fail loudly at the airport. It fails quietly on the shelf, as a stockout or a markdown.

Let me make it concrete, and I will state my assumptions out loud because I am not pretending to know your book. Suppose a category I buy carries a safety stock of 30 days. Suppose one third of that, 10 days, is genuine air-replenishment buffer, and the rest is sea-based. Suppose my exposure runs through a corridor whose outbound belly capacity is down 11.1%, matching what this airport reported. Holding my service level constant, the air-fed slice of my buffer has to stretch by roughly the same proportion: 10 days divided by 0.889 is about 11.25 days.

So my total safety stock moves from 30 days to about 31.3 days, call it 1.3 extra days of cover, and that spread applies across the whole category's working capital, not only the air-fed part. On a modest category that is real money in inventory and storage; on a fast-moving one it is a visible drag on sell-through. I would rather find that 1.3 days now, on my own terms, than discover it in August as a stockout.

That leads straight to how I would structure the award. The instinct when space looks uncertain is to grab a long-term contract and breathe. I do not think a straight long-term lock is right here, because the market has not settled enough for me to pay a premium for certainty I cannot verify. What I want is a split award with a hard trigger written in. Something like 70% of my corridor volume on a fixed long-term rate with a take-or-pay floor I can actually live with, and 30% left on spot so I keep a live read on the market. Then I attach a tripwire: if outbound belly capacity on the corridor falls another 5% below May's level, I re-tender the spot slice within 30 days instead of waiting for the annual review.

Multi-sourcing is the other lever, and it is the one buyers most often get wrong. Two suppliers on the same corridor is not multi-sourcing; it is the same bet with two names on it. What actually de-risks me is a second source on a genuinely different corridor, a different gateway, a different leg, a different failure mode. That costs me something, because splitting volume weakens my per-supplier leverage and can nudge unit prices up. But the alternative is a single-corridor dependency I now know can lose a tenth of its capacity without a phone call. In a category where a stockout costs me a customer, that premium is cheap. In a category where the product can tide over, I would not pay it.

The cash side deserves a line too. If I decide to carry more safety stock to cover a thinner air lane, that decision shows up as working capital, and working capital shows up in my payment terms. A stretched buffer and a relaxed payment schedule cannot both be true for long. So whichever way I move, more stock or a bigger split on spot, I have to re-run the terms conversation with the same rigour. Buyers who add buffer without touching terms are quietly borrowing from next quarter's balance sheet, and it always comes due.

I should say a word about how I would write the price itself, because a fixed number and a fixed number are not the same thing. On a corridor this jumpy, I would rather have an index-linked rate with a defined cap than a flat rate I have to renegotiate from scratch. The point of the index is not to predict; it is to remove the argument. If the reference moves, the rate moves, by a formula we both signed, up to a ceiling I can live with. What I will not do is accept an index with no cap, because an uncapped index on a corridor that can lose a tenth of its capacity in a quarter is not a hedge, it is an open position I did not mean to take.

Let me put the alternatives side by side, because none of them is free. Switching mode, air to sea, buys cost and spends time; for time-critical goods that is not a trade, it is a surrender. Re-sourcing to a supplier on a calmer corridor buys resilience but resets my qualification clock, my payment terms and often my unit price, and it takes months to make real. Holding more stock buys cover but ties up cash and raises obsolescence risk, which in fast-moving categories is the most expensive of the three. There is no free move. The only question is which cost I would rather book.

The traps are worth naming. One is reading stabilising as softening and walking into a negotiation to claw back price, only to find the space you wanted was promised elsewhere while you spent your credibility on a number that was never about price. Another is leaning on spot capacity you have not actually secured, on a corridor where the outbound leg is thinning; hope is not a booking. A third is letting the charge structure go unexamined, because on air moves what looks like a headline rate can carry a stack of charges whose basis does not match how your product actually moves; I would get the full build-up in writing, line by line, before I sign. And the quiet one is cultural: assuming that because a disruption happened once it stays contained. I would rather build a plan that survives a second shock than one that is optimal for the calm I am hoping for.

The cleanest way to hold this in my head is that a hub's cargo figure is a regional reading, and my lane is a lane. They overlap, but they are not the same object. A node can be down 12.75% while my specific city pair is tight, or flat while mine is loose. What transfers from the airport number to my ledger is not the percentage, it is the behaviour: capacity at conflict-adjacent nodes is now fast-moving, and my plans have to assume it. The percentage is the alarm; the behaviour is the lesson.

And let me be blunt about what I will not do with this number. I will not use a 12.75% decline to argue that rates on my lane should fall 12.75%; that is the kind of lazy linkage that gets a buyer walked out of the room. The number is a reading of one node's volume, not a price tag on my contract. If I walked in and demanded a proportional cut, I would be signalling that I do not understand the difference between volume and price, and any seller worth their salt would hold the line. The value of this piece of news is not in the percentage. It is in what it tells me about how fast the corridor can change under me, and that is what I actually take to the table.

One more thing I would nail down before signing, and it is the piece I cannot do myself. When I commit to a delivery promise to my own customer, the numbers I sign live or die on the last leg, not on the corridor. How tightly to watch that end-to-end commitment, which hour the window actually closes, which shipment to rescue first, Nadia Pryce knows that better than I do; my job is only to price the buyer's side of the bet. So my contract language has to hand off cleanly to that kind of operational discipline, or the promise I signed is worth less than the paper it is on.

Once it is signed, the work does not stop. I would put three numbers on a weekly one-pager and read them the same time every week: the outbound belly capacity on the corridor, the share of my volume riding a single gateway, and the gap between my planned and actual replenishment lead time. None of them needs a data platform to track; they need somebody who owns them. The value of a tripwire is not the threshold, it is that everybody agreed in advance what happens when it trips. Buyers who only look at these numbers after a problem already exists are looking at a stockout.

So here is what I would actually do, and it is not dramatic. I would not wait for a cleaner picture, because a cleaner picture is usually a more expensive picture. I would re-baseline my corridor exposure this week, find out honestly whether my own flow held while the node shrank, split my award with a written trigger, and push a second source onto a genuinely different road. The stabilisation in May is a useful read, but it is not a reason to relax; it is a reason to move while the noise is still low. If I were the buyer on this lane, that is the contract I would sign before the negotiation season hardens the price, and I would sign it knowing the one number I am really watching is the one nobody put in the headline.

↑ Back to the key points

— By Grace Yang