Freightos Air Index Holds 25% Above Pre-War as China-US Slips to $5.48/kg

The Analysis

Every corridor has a story and every corridor has a pothole, and this week the story is that one half of the map is draining while the other half is filling. The Freightos Air Index global benchmark is still roughly 25% above where it sat before the war, even after a 5% fall month on month. Read the lane detail underneath that headline and the picture stops being a single direction. China to United States sits at 5.48 dollars a kilo, seven percent below late February. South East Asia to Europe is back around 5.40. South Asia to Europe has come off its highs by ten percent to 4.60. South East Asia to North America is nine percent below peak at 6.41.

Then turn the map ninety degrees and everything I just wrote reverses. North America to Middle East gained six percent week on week to 4.57 dollars a kilo, and Europe to Middle East reached 3.87, still sitting near the top of its recent range. So the index fell five percent and the headline says soft, while the corridor that everything else in Eurasia depends on is going the other way at twice the speed. The same report carries container rates up sharply on the ocean side, which tells you the pressure is not confined to one mode.

I have spent enough years moving freight across this belt to know what that shape usually means. When the outbound blocks from East Asia soften and the Gulf strengthens at the same time, it is rarely because two unrelated markets made two unrelated decisions. It is usually because the same aircraft, the same crews and the same aprons are being pulled towards whichever lane currently pays, and the lanes that look unrelated on a rate sheet share a building on the ground.

Take the fact slice on its own terms before anybody interprets it. The global benchmark is about 25% above its pre-war level, and it got there after falling 5% in a month, so the level before that fall was something like a third above pre-war. China to United States at 5.48 is seven percent under late February, which puts the late February print somewhere near 5.89 on the same basis. South East Asia to Europe has returned to about 5.40. South Asia to Europe eased ten percent from its high to 4.60. South East Asia to North America is nine percent off peak at 6.41. On the other side, North America to Middle East added six percent in a single week to 4.57, and Europe to Middle East is at 3.87. Those are the numbers. Everything after this is what I think they mean on a map.

Why is the benchmark still a quarter above pre-war when the lanes everybody watches have been falling for weeks. Part of the answer is that the benchmark is a basket and the basket is not your lane, and part of it is that something structural moved into the air and has not moved back. The Red Sea diversion pushed a class of cargo out of the ocean network and into the air network, and that cargo does not come back on a schedule anybody can forecast. It comes back when the routing normalises, and until then it sits under the whole index like a floor. A 5% month on month fall is weather. The 25% is the floor, and floors are what you budget against.

Why now for the split. Capacity in the air is mobile in a way that no other mode can match, and it moves towards yield. Freighters follow the money within days, belly capacity follows passenger schedules which were filed months ago and cannot chase anything. So when the Gulf lanes firm up six percent in a week, the part of the fleet that can be redeployed gets redeployed, and it gets redeployed out of the lanes where rates are softening. That is the coupling nobody put in the headline: the softness on China to United States and the firmness on North America to Middle East are connected by the same metal.

Who feels it, and how much. High value electronics and pharma moving out of China and South East Asia are the winners in the soft lanes, and the saving is real: seven percent off late February on a lane that has spent two years being expensive is worth having. Low value e-commerce and apparel are not winners, because their economics were already marginal at these levels and a five percent move does not rescue a contribution margin. South Asia to Europe at 4.60 is the one to watch for the apparel and home goods crowd, because a ten percent give-back from the high is the difference between shipping by air and waiting for the next vessel for a lot of buyers in Dhaka, Karachi and Chennai. South East Asia to North America at 6.41 remains the most expensive of the big east-west blocks, and anyone quoting customers off that number is quoting off a lane that has not finished coming down.

Now the nodes, because this is where the money actually goes. Put your finger on the Gulf first. Dubai, Doha and to a lesser extent Istanbul are not just destinations, they are the transfer plates of the Eurasian air map. A very large amount of what nominally moves from Asia to Europe and from Asia to Africa physically passes across an apron in one of those places, gets broken down, gets rebuilt, and goes out again on a different aircraft. When local Gulf volume firms, the transfer cargo is the cargo that gets squeezed, and it gets squeezed for a reason that has nothing to do with your rate: transfer freight has to make a specific connection, local freight only has to make a flight. The ground handler optimises for the freight that is going to sit on its own apron tonight, and your connection is the one that gets rolled.

That is the point I have not seen anywhere in the coverage, so let me state it plainly. The 6% weekly gain on North America to Middle East is not a Middle East story. It is a story about your China to Europe transit time. If your eastbound routing relies on a Gulf transfer, the reliability of that routing is now being set by demand in a lane that appears on nobody's China-Europe rate sheet, and it is being set week by week while you are looking at a number that went down. The two lanes look unrelated on an index and they are the same set of loading docks. I have watched a perfectly priced transfer routing turn into a four day dwell because a market three thousand miles away firmed up, and the rate never moved to warn anybody.

The second node is South Asia, and it behaves differently. Delhi, Mumbai, Dhaka and Karachi feed Europe on a mix of belly and freighter capacity with fewer alternates than the China lanes have. South Asia to Europe easing ten percent to 4.60 sounds like relief, and for anyone booking today it is relief. The trap is that the alternates are thin, so when the lane firms again it firms fast, and the recovery tends to arrive in the same week that the apparel season needs the space. Book the softness if you need it, but do not build a 2027 sourcing plan on a number taken from the softest week of the year.

South East Asia is the third node and it carries a different risk. Ho Chi Minh City, Bangkok, Kuala Lumpur and Singapore have absorbed two years of diversification out of China, which means a lot of shippers who believe they diversified their risk are now queueing at the same four airports for the same fortnight of capacity. South East Asia to North America at 6.41 being nine percent off peak is not the same thing as slack: it is nine percent off a peak that was set by exactly that congestion. Look at absolute level, not at percentage off high. Six forty-one is still the most expensive block on the board, and the reason is ground handling capacity at the origin, not aircraft.

Now the choices, because a corridor is only useful if you can name the alternatives and price them. Three routes compete for the same Eurasian cargo and each has a bill attached.

Keep the direct. You pay the rate, you buy the transit, and you accept the volatility. Fine for anything with a hard date and a high value per kilo.

Move to a transfer through the Gulf or Turkey. You usually buy a lower rate per kilo because the transfer lane is priced to fill, and you pay for it in dwell, in a second handling, and in exposure to exactly the squeeze described above. The bill is not the rate difference, it is the day and a half.

Move to sea-air. Ocean to a Gulf or Mediterranean port, then air for the last leg. The rate can be dramatically better on paper, sometimes half the direct air number, and the transit is measured in weeks rather than days. It is a real answer for replenishment and a bad answer for anything with a launch date.

There is a fourth option and it never appears in an air index, which is why it is the one I get asked about most. The rail landbridge across Kazakhstan and the road corridors running through the Caspian and on into Turkey move the same Eurasian cargo across a different map. On today's numbers it sits between sea-air and direct air on transit, and well below direct air on rate. The appeal is obvious, and so is the catch: the landbridge has borders, and borders have gauges. Every gauge change is a lift of every box, and every lift is a place where your cargo stops moving while somebody else makes a decision. Add a week of dwell at a crossing to a routing that looked only two days slower than an aircraft and you have saved nothing at all. Use it for cargo with no date attached and a value per kilo that cannot justify air, and inspect the gauge change points before you inspect the price.

Changing the gateway is the cheaper version of the same question and it is where most people leave money on the table without noticing. On the South Asia lanes, booking out of Mumbai rather than Delhi, or trucking across to a coastal gateway with more freighter departures, can move a shipment two days and a few cents a kilo for the price of a domestic leg. That only works if the domestic leg is reliable, and in this part of the map it is reliable right up to the week it is not. I price gateway moves with a full day of slack and I do not promise the saving to a customer until I have run one.

The pothole in all of them is the same: every one of these routings concentrates at a small number of aprons, and the apron does not care which routing you booked. When it fills, the cargo that loses is always the cargo with the tightest connection and the lowest local priority.

Let me run the numbers so this stops being a matter of taste, and I will state every assumption. Assume 800 kilograms moving from east China to the United States east coast. Assume a direct air rate of 5.48 dollars a kilo, which is the published lane number, giving 4,384 dollars of freight. Assume an alternative routing via a Gulf transfer priced at 4.85 dollars a kilo, which is 11.5 percent cheaper and is my assumption, not a quote, giving 3,880 dollars. Assume transfer handling at 0.22 dollars a kilo, which is 176 dollars, and assume a dwell of one and a half days and two and a half days added to total transit once the second build-up is counted.

The gross saving is 504 dollars. Net of transfer handling it is 328 dollars. Against that you have two and a half days.

What are two and a half days worth. It depends entirely on what the cargo is doing. If this is replenishment for a selling line, the cost of the delay is gross margin you do not book. At a daily sales rate of 4,000 dollars and a gross margin of 30 percent, two and a half days of stockout costs 3,000 dollars of margin. Against a 328 dollar saving, the transfer routing loses by 2,672 dollars. If instead this is high value goods sitting in transit where the only cost of the extra days is working capital, at a total value of 240,000 dollars and a carrying cost of 8 percent a year, two and a half days ties up about 131 dollars. The transfer routing then wins by 197 dollars.

The break-even is clean and worth memorising. Divide 328 by 2.5 and you get about 131 dollars a day. Any shipment where one day of delay costs more than roughly 131 dollars of gross margin should not be routed through a Gulf transfer at these numbers. That is a low bar, and most selling lines clear it comfortably. The transfer routing is for cargo that is not in a hurry, or cargo that was going to the Gulf anyway.

Second calculation, on the floor rather than the lane. Assume your lane has moved broadly with the index, which is an assumption and not a fact. If 5.48 is about 25% above the pre-war level, the pre-war equivalent is roughly 4.38, so the structural premium is about 1.10 dollars a kilo. At 800 kilograms eight times a month, that is 6,400 kilos and about 7,040 dollars a month, roughly 84,000 dollars a year, that you are paying for a routing decision made by somebody else's navy. Waiting for February will not remove it. It comes off when the Red Sea routing normalises, and not before.

Where this could be wrong, because a corridor story always has a branch. The floor argument collapses if the diversion ends. If the ocean routing normalises, the cargo that moved into the air goes back into the ocean, the index loses its floor, and 25% above pre-war becomes ten percent above pre-war very quickly. That is the single biggest variable in this whole piece and it is not one you control.

The opposite branch is just as live. If Middle East demand keeps firming, the redeployed capacity stays redeployed, the eastbound lanes stop softening even without any change in eastbound demand, and the 6.41 and 5.48 numbers stop improving while your transit times quietly get worse. Watch the Gulf number as a leading indicator for your own lane, not as someone else's news.

The squeeze story has its own boundary condition and I will name it, because otherwise it reads as certainty. Transfer freight only loses priority while the local market is paying more than the transfer market is. If the Gulf lanes roll over, the ground handler's optimisation flips back within a week or two, and your connection becomes the profitable one again. So the honest reading is not that Gulf transfers are broken, it is that they are currently priced for a condition that can reverse without warning, and the way to hold that risk is to buy the buffer in the promise you make rather than in the rate you pay.

One more branch. Belly capacity keeps recovering as passenger flying keeps rebuilding. Every point of belly recovery is supply that arrives on a schedule nobody in cargo controls, and it lands hardest on exactly the lanes that look tightest today. If you are signing anything longer than a quarter, leave fuel and security surcharges floating and cap only the base, so that recovery reaches your invoice instead of your forwarder's margin.

So, does this road still get walked. For high value, time sensitive cargo out of China and South East Asia, yes, and today is a better day to book than last month was. For transfer-dependent routings across the Gulf, walk it with your eyes open and a buffer written into the promise you make your customer. For anything low value and replenishment driven, the air number is not your problem, the two and a half days is your problem, and the honest answer is that this cargo should not be on an aircraft at all.

↑ Back to the key points

— By Ömer Kaya