Global air cargo spot rates eased 3% month on month in August to an average of $3.13 per kilogram, though they remain 24% above last year and the pace of year-on-year growth has slowed for three straight months, per Xeneta data cited by J.M. Rodgers. Demand was still up 6% year on year in August, with supply growth lagging. Shippers are increasingly buying short-term capacity instead of locking contracts, while AI-related shipments keep transpacific pricing elevated into Golden Week.
Supply Chain Action Points
What this means for your business — and what to do about it:
Air cargo pricing has moved into a phase where the spot market and the contract market no longer point the same way. Xeneta data cited by J.M. Rodgers puts the August global air cargo spot average at US$3.13 per kilogram, down 3% month on month but still 24% above August 2025, and the year-on-year growth rate has now narrowed for three consecutive months.
Demand is not the problem. August volumes were up 6% year on year and supply growth is still lagging, which is why a 3% month-on-month dip has not turned into a real correction. What has changed is behaviour: shippers are buying short-term capacity rather than committing to long-term contracts, which pushes the pricing risk onto whoever books last.
The consequence shows up in two places. Transpacific rates remain elevated on the back of AI-related shipments and are expected to tighten again before Golden Week, and the pressure is spilling into surface transport, with the US LTL rate index up 4.5% in August. The five lists below turn those numbers into decisions for exporters, cross-border e-commerce sellers, factories, brands and procurement teams.
For Exporters
Xeneta data cited by J.M. Rodgers puts the August global air cargo spot average at US$3.13 per kilogram, 24% above the same month in 2025. If you are still issuing air quotations that were built a year ago and simply rolled forward, that quotation is now roughly a fifth short of the market. The 3% month-on-month decline is no reason to relax your rate clauses, because the year-on-year growth rate has narrowed for three straight months, which describes a market plateauing at a high level rather than one heading back to 2025 prices. Trade terms decide who absorbs that gap: quote FOB and the buyer nominates the forwarder and carries the 24%; quote CIF or DDP and the exposure runs from origin to destination on your own account.
Work the exposure through one lane rather than across the whole book. Assume your team ships 30 air consignments a month at an average chargeable weight of 800 kilograms each, and assume the 3.13 dollar August average is representative of your mix. That is 800 multiplied by 3.13, or US$2,504 per consignment, and roughly US$75,120 a month at today's spot. Rebuilding last year's level from the stated 24% gap implies an August 2025 rate of about US$2.52 per kilogram, or US$2,016 per consignment. The difference is US$488 per consignment and about US$14,640 a month. That is the size of the hole left by a quotation without a rate-review clause, and it is also the figure your clause parameters should be designed against.
Put a date and an owner on the next round. From 15 September, every open air quotation older than 14 days is re-priced before confirmation, and transpacific validity is capped at seven days while Golden Week demand is still accumulating. Assign the rate clause to the sales lead rather than the logistics coordinator, because margin is decided at the moment of quoting, not at the moment of booking. Then write an automatic trigger: if the published benchmark for that lane moves more than 3% against the quoted rate before departure, the quotation is withdrawn and re-issued rather than absorbed internally. A clause that nobody can invoke without an argument is not a clause.
Three alternatives are worth pricing side by side. The first is moving 20% to 30% of volume to ocean where transit tolerance allows, trading speed for a much lower cost base. The second is routing through a second gateway to arbitrage the regional spread, which adds drayage and can complicate origin documentation. The third is buying a partial capacity guarantee on the single lane carrying your key accounts, trading flexibility for a price floor. The most common mistakes are procedural. Treating the trade term as a formality means discovering at invoice stage that CIF or DDP made the freight exposure yours. And an origin declaration or certificate of origin that does not match the actual routing can hold a shipment long enough that everything the rate clause saved is spent on storage and re-handling.
- From 15 September, re-price any open air quotation older than 14 days before confirmation; cap transpacific validity at 7 days. Owner: sales lead.
- Write an automatic trigger into every quotation: benchmark movement of more than 3% against the quoted rate before departure forces re-issue.
- Hold FOB-lane air cost exposure below 10% of monthly air spend, measured against the US$75,120 baseline (30 consignments x 800kg x US$3.13/kg).
- By 30 September, price an ocean or second-gateway alternative for every lane able to tolerate 5 or more transit days. Owner: export manager.
- Audit the top 10 air lanes so declared origin, commercial invoice and AWB agree; name one compliance owner per lane.
For Cross-Border E-commerce
A US$3.13 per kilogram August spot average, up 24% year on year, lands directly on cross-border e-commerce landed cost. Assume your restock profile averages 0.5 kilograms net weight per unit. Air freight then costs US$1.57 a unit, against roughly US$1.26 a year ago on the same 24% gap, so unit cost is about 30 cents higher. Thirty cents looks survivable until it meets a low-price SKU, where it can wipe out the entire contribution margin. The surface leg is moving the same way, with the US LTL rate index up 4.5% in August, so the final mile is not a shelter from the air market.
Take a month of restocking. Assume 40,000 units a month at 0.5 kilograms net each, which is 20,000 chargeable kilograms. At US$3.13 per kilogram that is US$62,600 a month; on the same basis a year ago it was about US$50,480. The gap is roughly US$12,100 a month, or close to US$145,000 across a year. Note what that does to your pricing options. Market demand grew 6% year on year, so you cannot assume volume growth will dilute the increase. If half of that restock is low-margin product, a 5% retail price increase does not recover a 30 cent unit cost increase, because the percentage comes off a smaller base than the cost increment does.
Move the plan onto replenishment rhythm. By 1 October, shift A-class best-sellers from a fortnightly cycle to twice weekly, halve the batch size, and cut safety stock from 21 days to 14 days so frequency replaces inventory. Lock the top 20 SKUs into a supply-protection list and accept stock-outs only outside that list. Rebuild the last-mile layer in the same pass instead of looking only at the head haul: convert the 4.5% LTL increase into a per-parcel cost using your August volumes and put that number into the late-September price review. Anyone who re-prices only the air leg will discover in November that the domestic leg was the second half of the increase.
Three alternatives exist. Low-value bulky SKUs move to ocean LCL or to overseas warehouse pre-positioning. High-value time-sensitive SKUs stay on air. The middle group splits across two nodes, with half the inventory sitting closer to the consumer. The trade-off is capital and obsolescence risk: over-filling an overseas warehouse ties up cash rather than freight cost, and slow-moving stock does not become cheaper by sitting abroad. Two mistakes recur. The first is counting only the head-haul air rate and ignoring the stacked domestic increase. The second is treating the return rate as a constant: once return handling stretches, the resale window closes and a returned unit stops being a unit you can sell.
- By 1 October, move A-class best-sellers from fortnightly to twice-weekly replenishment, halve batch size, cut safety stock from 21 to 14 days. Owner: merchandising.
- Lock the top 20 SKUs into a supply-protection list; stock-outs are permitted only outside that list.
- Convert the 4.5% LTL increase into a per-parcel cost from August volumes and complete the price review in late September. Owner: pricing.
- Cap head-haul air spend at US$62,600 a month (40,000 units x 0.5kg x US$3.13/kg); anything above that diverts to ocean LCL.
- Clear returned units within 24 hours of arrival at the warehouse; named owner: warehouse supervisor.
For Manufacturing Plants
August demand was up 6% year on year while supply growth lagged, pricing is still 24% above last year, and the transpacific squeeze is expected to tighten again before Golden Week. For a factory this is not a freight cost question first; it is a line-stop question. Two days lost on a key component's inbound leg is enough to invalidate a production schedule that took two weeks to build. The US$3.13 per kilogram spot level is what decides which components are worth flying and which must simply be carried on inventory. It also sets the number your safety-stock policy should have been written against, because a policy written at last year's rate understates the cost of the buffer it exists to avoid.
Run the arithmetic on consumption. Assume one line consumes 12,000 key components a month at 0.15 kilograms each, giving 1,800 chargeable kilograms. At US$3.13 per kilogram that is US$5,634 a month; on the same 24% basis a year ago it was about US$4,544, so the increase is roughly US$1,090 a month. That is small enough to look ignorable, which is exactly why it gets mis-decided. Send it by ocean instead to save that cost, and on the assumption that the alternative adds 18 days of transit, the line needs 18 days of extra consumption in stock, which is 7,200 more units of working capital and warehouse space. The saving is US$1,090 a month; the cost is a stock position, not a freight line.
By 1 October, raise A-class spares and key components to 45 days of consumption and hold B-class at 30 days. Any item below 30 days of cover triggers air replenishment immediately with no approval queue, because the queue is where the line stops. Make the equipment engineering manager, not the buyer, accountable for line-stop risk, because that role can price a stoppage and the buyer cannot. Book Golden Week capacity early rather than late: complete booking for the 1 to 8 October window by 25 September. Capacity is being bought short-term across the market, which means the late booker pays the spot rate and carries the space risk.
The alternatives are dual-sourcing, rerouting and inventory. Dual-source key components so a second supplier is already qualified even at a 5% to 8% unit premium, because an unqualified second source is not a backup. Reroute part of the air volume through a second gateway to capture a regional spread. For components genuinely not worth flying, buy time with safety stock and accept the carrying cost. The traps sit in documentation and classification. Spare parts containing batteries or magnets are refused at acceptance if the dangerous-goods determination was not done in advance. An AWB whose description does not match the customs declaration produces a shipment you cannot collect. And the most common error of all is reading the market's 6% demand growth as your own order growth.
- By 1 October, raise A-class spares and key components to 45 days of cover and B-class to 30 days; below 30 days triggers air replenishment without approval.
- Complete booking for the 1-8 October Golden Week window by 25 September; no late booking after 30 September.
- Name the equipment engineering manager as line-stop owner; qualify a second source for every key component, accepting up to an 8% premium.
- Hold key-component air spend at US$5,634 a month (12,000 units x 0.15kg x US$3.13/kg); the excess moves to ocean with matching inventory.
- Complete dangerous-goods determination and reconcile the AWB description against the customs declaration before any spare part flies.
For Brand Owners
A US$3.13 per kilogram spot average and 24% year-on-year growth translate into something more concrete than a freight line: the delivery promise you make to customers now costs materially more than last year's model assumed. Demand is up 6% year on year, capacity growth is lagging, and the transpacific lane is expected to tighten further before Golden Week. Meanwhile the US LTL rate index rose 4.5% in August, so the final mile is not insulated either. Every day you pull off the promised delivery window costs margin that has to be recognised at the pricing stage rather than explained away afterwards. A promise priced wrong does not appear as a logistics variance; it appears as a margin miss.
Model the direct-to-consumer book. Assume 2,000 cross-border direct orders a day, of which 30% fly, at 0.6 kilograms net per unit. That is 360 chargeable kilograms a day, which at US$3.13 per kilogram costs about US$1,127 a day, or roughly US$33,800 a month. On the same 24% gap, those 360 kilograms cost about US$908 a day a year ago, so the increase is about US$219 a day and roughly US$6,570 a month. That covers only the air segment. Add the 4.5% LTL increase on the domestic leg before deciding whether the current delivery promise can stay unchanged.
Reset the published promise before 30 September. State the promised number of days for transpacific air orders and specify that the count is in working days. On lanes where you cannot hold the promise reliably, widen it by one day rather than relying on customer service to explain delays after the fact. Write the channel priority down explicitly: bonded and overseas warehouses protect best-sellers and the top 20 high-margin SKUs first, while platform warehouses and store replenishment queue behind them. In the 3PL contract, put three numbers in writing: the committed days, the threshold below which no compensation is owed, and the compensation amount above it.
The alternative is tiered fulfilment. High-margin orders keep a fast promise on air. Mid and low-margin orders default to ocean plus overseas warehouse pre-positioning with a 7 to 10 day tolerance added. During promotions only selected SKUs retain the fast promise. The cost is visible inconsistency, because customers see different arrival dates for different items in the same store, which requires matching page copy. Three mistakes recur: a promise that states days without stating working days; a returns window measured from dispatch instead of delivery; and fragmented information where the supplier, the 3PL and the internal system each hold a different timestamp and tracking number, so that when something breaks nobody can say where the goods are.
- By 30 September, republish the transpacific delivery promise with working-day wording and a stated compensation rule. Owner: brand operations.
- Set channel priority: bonded and overseas warehouses protect the top 20 high-margin SKUs; platform warehouses and stores replenish after them.
- Add three numbers to every 3PL contract: committed days, the no-compensation threshold, and compensation per breach.
- Cap D2C air spend at US$33,800 a month (2,000 orders/day x 30% x 0.6kg x US$3.13/kg); the excess moves to overseas warehouse pre-positioning.
- Launch tiered fulfilment: fast air promise for high-margin orders, 7-10 extra days for mid and low margin, with page copy updated the same day.
For Procurement Teams
The market has changed behaviour, not just price. Shippers are moving away from long-term contracts and buying short-term capacity instead, the August spot average slipped 3% month on month to US$3.13 per kilogram while staying 24% above last year, and year-on-year growth has narrowed three months running. For procurement that combination means the value of a contract is not in locking today's number; it is in locking a relative cost after the growth rate has already turned. With demand up 6% year on year and capacity lagging, negotiating closer to Golden Week means negotiating with less leverage. Everything below assumes you have authority over the air capacity line and a mandate to move the ratio between contract and spot.
Run the annual exposure. Assume 600 tonnes of air freight a year at the August spot of US$3.13 per kilogram, which is 600,000 kilograms and about US$1.88 million. On the same 24% basis, last year's equivalent was about US$1.51 million, so the year-on-year gap is roughly US$363,000. That gap is the negotiating budget and belongs at the top of your brief. If a contract can hold your annual cost inside half that gap, the contract is worth signing. If the carrier offers only a 3% one-sided discount, the discount is worth less than the flexibility you give up, and keeping a larger spot share while the year-on-year curve flattens is the better trade.
Fix the timetable. Complete the contract-versus-spot 60:40 modelling by 20 September, and settle the index-linked element by 25 September: the BAF reference index, the reference period, the adjustment frequency, the notice period and the cap. Capacity diversification belongs on a route list rather than a supplier list, so keep at least one backup forwarder per main lane and prove it with one live shipment before 1 October, because an untested backup is an assumption. Then add three clauses: the trigger threshold and notice period for rate adjustment, a minimum quantity the carrier must guarantee by lane, and explicit force majeure wording with a named alternative arrangement.
Three structures are available: all spot, all contract, or a contract base with spot for peaks. The trade-off is certainty against optionality, and it should be decided lane by lane rather than across the whole book, because lanes differ in volume predictability. The traps are in the wording. A contract that locks the rate but not the space still gets rolled in a tight market. An index-linked clause whose reference does not match your actual routing rises when the market rises and does not fall when the market falls. And force majeure wording that does not cover Golden Week, typhoon season or a destination port strike is discovered to be useless at the only moment it mattered.
- By 20 September, complete contract-versus-spot 60:40 modelling and present both annual cost outcomes to the business.
- By 25 September, settle the BAF index-linked clause: reference index, reference period, frequency, notice period and cap. Owner: procurement lead.
- Keep one backup forwarder per main lane and complete one live shipment validation before 1 October.
- Use US$1.88m as the annual baseline (600 tonnes x US$3.13/kg) and US$363,000 as the year-on-year gap that sets the negotiation ceiling.
- Add to contract: minimum guaranteed space by lane, a 3% rate-adjustment trigger with a notice period, and force majeure covering Golden Week, typhoons and port strikes.