Air cargo enters early peak as China-US rates hit $6/kg with volumes up 13% year on year

Air cargo has walked into its early peak, and the rates are moving the way they do when new electronics and AI hardware hit the tarmac at once. The Freightos Air Index on 28 September 2026 put Greater China and Asia to North America at six dollars and three cents per kilo, to Europe at four dollars and seventy-one cents per kilo, and the global all-lanes rate at three dollars and four cents per kilo. China and Hong Kong to the US spot sat around six dollars a kilo, up twenty-six percent year on year, with volume up thirteen percent; the TAC air index was up nineteen percent year on year. Guangzhou and Shenzhen to Los Angeles ran thirty-six to thirty-eight yuan per kilo, to Chicago forty-one to forty-two, to New York forty-two to forty-three.

Demand is being led by the new consumer electronics cycle, AI servers, and semiconductors, which are exactly the goods that ride passenger-belly and freighter capacity at the same time and crowd each other out. When a fresh phone launch and a rack of AI servers both need to fly in the same week, the capacity that was loose in late summer tightens fast, and the rate follows. This is not a vague seasonal lift; it is a specific, product-driven squeeze where the cargo that pays best gets the space and everything else waits or pays more.

Read the index spread carefully. The China and Hong Kong to US spot at six dollars a kilo, up twenty-six percent year on year, is the sharpest move, and it lines up with the TAC index up nineteen percent across the board. The gap between the two tells you the US lane is heating faster than the global average, which makes sense because US-bound electronics and AI kit ship in the highest volume and the dollar value justifies the premium. If your freight is US-bound out of China, you are in the hottest lane and should act first.

The China and Hong Kong to Europe leg at four dollars and seventy-one cents is cheaper than the US lane but still elevated, and it is the one where a lot of the AI server and semiconductor volume also flows. Europe is not the bargain it was; it is just the less-hot of two hot lanes. For an exporter serving both markets, the pricing signal is to prioritize US space bookings because that is where the squeeze and the rate are worst, while Europe can be planned a touch later but not ignored.

The Guangzhou-Shenzhen numbers are the ones to watch if you are a South China shipper. Thirty-six to thirty-eight yuan per kilo to Los Angeles, forty-one to forty-two to Chicago, forty-two to forty-three to New York tells you the transcontinental US move costs a clear premium over the coastal gateway. Chicago and New York are inland destinations that need a second leg, and that leg is where the per-kilo number jumps by five to seven yuan. If your US delivery is to the interior, budget the full forty-one to forty-three, not the LA number, or your margin math lies to you.

Volume up thirteen percent year on year alongside rate up twenty-six percent is the classic early-peak signature: more goods chasing the same planes, and the rate rising faster than the volume because capacity is not growing in step. Capacity tightens in late September, which is the window the index is catching. The warning to shippers is plain: lock space and price early, and list peak surcharges separately so the customer sees the real number and you are not left holding a spread between the rate you quoted and the rate you paid.

For an importer-exporter the action splits by direction. On the import side, if you are bringing components or finished electronics from Asia by air because the ocean lane is too slow for the launch window, your air cost just became a real line in the margin that was not there a year ago at this level. On the export side, if you are flying goods out of China to meet a US or European deadline, you are competing with every other electronics and AI shipper for the same bellies, and the one who books early and pays the listed peak gets the box moved.

Locking space early is the single highest-value move right now. With capacity tightening in late September, a booking made today at a named rate beats a booking made next week at a higher rate and a worse slot. The carriers and forwarders who hold freighter capacity allocate it to committed volume first, so a standing agreement with a peak block gets you space when the spot buyer is rolled. If you have not already, open the peak block conversation with your freight forwarder this week, not after a shipment misses.

Listing peak surcharges separately is the discipline that protects your customer relationship. A single all-in air quote that hides a security, fuel, or peak fee invites a fight when the customer compares it to a competitor who broke it out. Show the base rate, the fuel, and the peak as distinct lines, and the customer understands the market instead of suspecting you of padding. Transparency on a six-dollar kilo lane is what keeps the account when the rate is already hard to swallow.

The AI server and semiconductor flow deserves its own planning because it is not like apparel. Those shipments are high value, often time-critical, and sometimes under export control paperwork that slows handling. When the lane is tight, the high-value urgent box usually wins the space, but the paperwork-heavy box can still miss if the forwarder is slammed. Build the extra handling time into your plan for controlled goods, and do not assume a tight market will bend the process for you.

Chicago and New York as interior US points need a capacity buffer of their own. The second leg from LA or the coastal gateway to the interior is a separate capacity pool, and late September tightness hits it too. If your delivery is Chicago at forty-one to forty-two yuan per kilo, confirm the onward leg is booked with the first, not assumed, because a box that clears LA and then waits a week for the interior flight is a box that missed its promise. Treat the transcon as part of the same booking, not a footnote.

A worked example. Say you fly five hundred kilos of new consumer electronics from Shenzhen to Los Angeles at thirty-seven yuan per kilo, that is eighteen thousand five hundred yuan of freight. The same five hundred kilos to New York at forty-two yuan is twenty-one thousand yuan, a two thousand five hundred yuan difference for the interior leg alone. If you quoted the LA number to a New York customer, you lose two thousand five hundred yuan on the shipment, and at peak when capacity is tight that interior leg is exactly where the rate spikes first. The geography in the quote has to match the geography of the delivery.

The global all-lanes rate at three dollars and four cents is the one number that should not fool you. It is an average across every lane, and it sits well below the China-US and China-Europe numbers because it is pulled down by lanes that are not in peak. Using the global average to price a China outbound shipment is how operators underquote and then eat the difference. Always price air off the lane-specific index, never the blended average, during a peak like this.

Watch the late-September capacity turn. The note that capacity tightens late in the month is the trigger: as October approaches, the early-peak demand meets the pre-holiday push and the lane can snap from tight to slammed. The shippers who locked in September space and separated surcharges are fine; the ones who waited for a softer rate are the ones calling for emergency capacity at the worst moment. The pattern repeats every year and every year the waiters lose.

For the importer specifically, the early peak is a reason to pull forward any air-dependent order you can. If a component has to fly to keep a production line or a launch alive, bring the PO forward into the locked space now rather than betting the line on a rate that is climbing. The cost of early air is known; the cost of a stopped line while you hunt for capacity is not, and the second is always larger.

For the exporter, the early peak is a reason to communicate now. Tell your US and European customers that air rates are up twenty-six percent year on year and capacity is tightening, so their deadlines need bookings this week, not the week before flight. A customer who knows the lane is hot plans around it; a customer who finds out at the dock argues about why the rate moved. The call you make in September is cheaper than the apology you make in October.

So the rule for this air peak is straightforward. Lock space and price early on the lane you actually use, list every surcharge as its own line, book the interior US leg with the first, plan extra time for controlled goods, and price off the China-specific index rather than the global average. The early peak is here, the electronics and AI volume is real, and the operator who treats six dollars a kilo to the US as a reason to move now rather than wait will keep the boxes flying while the waiters explain why theirs did not.

Choosing freighter versus belly capacity is the first call when the lane tightens. Passenger belly is cheap and plentiful in normal times but it disappears when airlines trim flights or when the hold is full of someone else's urgent electronics. A freighter block you hold with a forwarder keeps space when bellies clamp shut. In an early peak led by launches and AI kit, the freighter is the insurance, and the cost difference is small next to a missed launch.

Book through a forwarder who holds main deck, not just the carrier's app, because the forwarder can blend your volume across several freighters and several airlines to find the slot a single carrier cannot. The China to US lane at six dollars a kilo is competitive for space, and a forwarder with multiple relationships gets you moved when the one-airline shipper waits. The per-kilo saving from blending beats the convenience of a single booking screen.

Density and volume weight decide your real per-kilo cost more than the rate sheet admits. Air freight is charged on the higher of actual and volumetric weight, and electronics can be light or dense depending on packaging. If your AI servers pack tight, you pay by actual weight and the rate stings less; if your consumer electronics ship in bulky cartons, volumetric weight quietly doubles what you owe. Re-pack to cut volume before you complain about the rate, because the index does not care how you box.

Build a rate card per lane and review it weekly through the peak, because the six-dollar US and four-seventy-one Europe numbers are moving targets. A card that lists LA, Chicago, New York, and the European gateways with their current per-kilo and per-yuan figures lets you quote a customer in minutes without guessing. The operators who price from memory in a peak like this are the ones who underquote on Monday and eat the gap by Friday.

Document readiness is the unglamorous reason boxes miss in a tight lane. AI servers and semiconductors often need export licenses, end-user statements, or dual-use paperwork, and a slammed forwarder will prioritize the clean box over the one waiting on a missing form. Have every license and statement in the file before you book, not after, because a held shipment pays the peak rate and still sits on the floor. The paperwork is part of the capacity you compete for.

A communications template to your customer saves more than it costs. A short note that air rates to the US are up twenty-six percent year on year, capacity tightens in late September, and we are locking space this week with peak surcharges shown separately, asks the customer to confirm the booking now rather than debate later. The customer who knows the lane is hot plans around it; the one surprised at the dock argues about why the number moved, and that argument is the one you lose.

Plan for the rate to keep climbing, not to fall. The early peak with volume up thirteen percent and rate up twenty-six percent has room to run if the electronics and AI demand holds, and a lane that snaps from tight to slammed in late September can print seven dollars a kilo to the US before October. Lock what you can at six and treat any unbooked volume as a seven-dollar risk in your model, because hoping for a dip while demand is rising is how you get stuck at the top.

A second worked example, this time to Europe. Say you fly three hundred kilos of semiconductor samples from Hong Kong to Frankfurt at four dollars and seventy-one cents per kilo, that is fourteen hundred thirteen dollars of freight, plus fuel and peak listed separately. If the same samples went to Chicago instead, the US lane at six dollars would be eighteen hundred dollars for the same weight, a near four-hundred-dollar gap that argues for routing critical but less urgent volume to Europe when the US lane is slammed. The lane choice is a cost lever the rate spread hands you for free.

Keep a standing peak block even if you hate committing, because the alternative is the spot market at its worst moment. A block of a few hundred kilos a week at a named rate with a named slot means your launches fly; without it, you beg for space the week everyone else launches. The block costs a commitment, but a commitment you would ship anyway is cheaper than emergency capacity at the peak. The waiters pay double; the committed fly.

The disciplined close is that air in this peak rewards the operator who moves first and shows the math. Lock space on the lane you use, break every surcharge into its own line, pack for actual weight, clear the paperwork before booking, and price off the China-specific index. The electronics and AI volume is real, the capacity is tightening, and the box that is booked and documented flies while the one still being quoted sits on the floor.

Treat the index as a trigger, not a target. The Freightos and TAC numbers are where the market is, not where it will be next week, and in a peak led by product launches the move is often abrupt. Set a rule with your team that any China outbound booking not locked within forty-eight hours of a rate quote is re-priced at the current index, so a slow internal approval does not leave you holding last week's number against this week's market. The discipline of re-pricing fast is what stops the gap from opening.

Consider a freighter charter split for the highest-value, most time-critical lanes if your volume justifies it. When the US lane prints six dollars and spot space is scarce, a partial charter that fills the rest of the aircraft with other shippers' goods can lock your critical boxes at a known rate while the market moves. This is not for everyone, but for an importer-exporter moving AI servers in volume, a shared charter is the difference between flying on schedule and begging for a belly. The threshold is lower than most think once the peak bites.

Do not let the peak surcharge surprise your finance team. Because the peak is listed separately on the quote, your margin recognition has to match it, and a shipment booked at a base rate with a peak added at invoice can distort a month's freight spend if finance expected the base only. Brief the finance desk that the all-in air cost this quarter includes a peak line that did not exist a year ago, so the budget variance is explained before the close, not argued after it. A surprise in your own books is the cheapest surprise to prevent.

Review your forwarder's performance weekly against the rate card, not just at quarter end, because in a peak the gap between promised and delivered space shows up fast. A forwarder who quoted you a slot and then rolled your box when the lane snapped owes you the peak difference or a credit, and the only way to catch it is to watch the bookings against the card every week. The operator who audits the peak as it happens protects margin; the one who discovers the roll at the close discovers the loss too late to recover it.

• Lock air space and rate now on the specific lane you use, especially China to US at six dollars a kilo, before late-September capacity tightens further.
• List peak, fuel, and security surcharges as separate lines on every air quote so the customer sees the real market number and you avoid a margin fight.
• Book the US interior leg, Chicago or New York, together with the first flight rather than assuming onward capacity, because the transcon is a separate tight pool.
• Open a peak block agreement with your freight forwarder this week so committed volume gets allocated freighter space before the spot buyer is rolled.
• Price every China outbound shipment off the lane-specific Freightos or TAC index, never the three-dollar global all-lanes average, during this early peak.