TD Cowen/AFS Index: Parcel, LTL and Truckload Costs All Set Q4 Records on Fuel and Capacity

The Analysis

A carton leaves a transload shed off the 60 in Ontario, California, sometime after two in the morning. It has already crossed an ocean, a container terminal at Long Beach, a dray yard, and a distribution centre floor, and every one of those crossings carries a price somebody sat down and argued over. The last crossing costs more than most of them and nobody argues about it at all. It is the five digits at the end of an address. On 4 January 2027 FedEx moves 2,533 ZIP codes onto its remote list at $17.75 a package, a move that has been described as doubling what a residential stop costs in the surcharge column. The box will not travel one inch further on the 4th than it did on the 3rd. The border moved underneath it.

I have worked this lane. Twenty years of chasing freight across three continents teaches you one thing about where the money goes: the node that bites you is almost never the one with the crane on it, it is the classification node, the desk where somebody decides what your cargo is called. A surcharge table is exactly that kind of node. It looks like administration. It behaves like a toll booth on a road you already paid to build.

The Q3 TD Cowen/AFS Freight Index landed this week with parcel, LTL and truckload costs all at record highs for the fourth quarter of 2026. LTL rate per pound is 82.5% above its 2018 baseline. Truckload rate per mile is 23.4% above the same baseline and 14.7% higher than a year ago, the steepest year-on-year jump since the first quarter of 2022. And the number I would take into a room: the fuel surcharge has climbed 22.25 percentage points since the first quarter of 2021, and 72% of that climb came from carriers redrawing their own tables, not from diesel. At $2.91 a gallon the surcharge would still sit at 23.75%, three times where it stood in 2021. Do not shrug off this crossing. Two and a half thousand ZIP codes are about to change nationality, and the only shippers who will know in time are the ones who ran the map in November.

Start with what is genuinely load-bearing, because the headline number is not the one that matters. The index says all three modes set records in the fourth quarter. Inside that, three figures carry weight, and a fourth explains them. LTL rate per pound sits 82.5% above its 2018 baseline. Truckload rate per mile is 23.4% above baseline, up 14.7% year on year, the sharpest move since the first quarter of 2022. Parcel is at a record on the back of a January reclassification of 2,533 ZIP codes at $17.75 a package. Then the figure that turns the whole story: 72% of the 22.25-point rise in the fuel surcharge since the first quarter of 2021 came from carriers redrawing the table, not from the price of diesel.

Sit with that for a moment, because it is the difference between a cost and a price. Diesel at $2.91 a gallon is not a fuel crisis; it is roughly the middle of the road. Yet the index says the surcharge would still be 23.75% at that price, three times its 2021 level. A surcharge that keeps three-quarters of its climb while the commodity it is named after goes flat is not tracking the commodity. It is tracking something else, and the index has told you what: the design of the table.

Why now, and why at this level? The mechanism is not mysterious if you have sat on the buying side of a bid in the last three years. Base rates were competed down hard when capacity was loose, and the money handed back in the base rate did not disappear, it moved sideways into columns that were not on the bid sheet. A surcharge table is the easiest place in any contract to move money sideways, because it arrives dressed as arithmetic rather than as a price. Nobody argues with a formula. Everybody argues with a rate. That asymmetry is worth more to a carrier than three points of base rate, and the index is telling you it has been used.

Now add the capacity side, because it is why the timing works. Truckload rate per mile rising 14.7% year on year, the fastest since the first quarter of 2022, is not a fuel story either. That is the signature of a market where trucks have left and the ones left are choosing their loads. When a carrier has pricing power on the linehaul, it also has cover to redraw the columns underneath it. Q4 makes it stick: peak volume is on the road, nobody is going to walk away from a carrier in November, and the contracts that reset on 1 January will be signed during exactly the weeks when the index is at its record.

The FedEx piece fits the same logic and deserves its own paragraph, because the date is doing real work in it. A remote classification is how a network charges for the tail of its own footprint: the last twenty kilometres into a place where the density is thin. Reclassifying 2,533 ZIP codes does not change a single route or add a single mile of driving. It changes what those miles are called. Putting it on 4 January, the first working days of a new contract year and the week after peak, is the calendar equivalent of changing the toll after the queue has formed. There is nothing improper about it. There is also nothing accidental about it.

Who eats this, and how deep? Closest to the blast are residential shippers with an exurban or rural footprint, and they should check their map before anything else, because $17.75 a package is not a rounding error on a low-value parcel; it can be the whole margin. Inside that group the exposure is wildly uneven. A shipper whose residential base is dense suburban ZIPs inside a metro will lose almost nothing. A shipper selling into the outer ring, into lake country, into the kind of ZIP where the driver does eleven stops in an hour instead of thirty, will discover that a slice of its book just got repriced by a stranger.

LTL comes next. At 82.5% above its 2018 baseline, the effective price of a pound is now high enough that the classification of that pound matters more than the discount on it. Freight class, density, re-weigh practice, and whether your packaging has quietly got lighter or bulkier over two product cycles: these decide the bill more than the negotiated rate does. A shipper who has not re-weighed its top lanes in two years is paying last decade's freight class for this decade's carton.

Truckload comes next and splits in two. Contract truckload, bid annually, absorbs the 23.4% above baseline slowly and with warning. Spot truckload absorbs it in a single phone call, and the 14.7% year-on-year move is a spot-market number in spirit even where it shows up in contracts. Anyone running a meaningful share of spot on lanes longer than a day's drive is exposed to the whole move at once, usually in the fourth quarter, usually on the freight that has to move.

One more group can close the file right now: shippers whose volume is commercial, urban and dense, whose freight is palletised rather than parcelised, or who run their own fleet. If that is you, the January reclassification touches almost nothing and the surcharge argument is somebody else's fight. Read the rest for the LTL and truckload numbers and stop there.

Timing is where this becomes a job of work rather than a reading exercise. There is one hard date: 4 January 2027. Everything else is a lead time measured backwards from it. If you want to know how many of your parcels land in the 2,533 affected codes, you need that answer in November, because the things you can do about it each take weeks. Moving inventory between fulfilment nodes takes weeks. Re-pricing a residential delivery promise on your own website takes weeks of merchandising and customer-service work. Consolidating to commercial addresses, or to a pickup-point network, takes a commercial conversation with customers who currently expect a box on the porch. None of those start on the 4th of January. They start the day you run the map.

The other clock is the budget. Q4 records in an index published in late October land squarely in the weeks when 2027 freight budgets are being closed. A budget built on this year's effective rate is a budget built on a table that is about to be redrawn again, and a table that has already been redrawn once without diesel's help.

Here is the part the headline left out, and it is the reason I sat down to write this. The fuel surcharge has stopped being a hedge against fuel and become an instrument for margin, and that changes what you should be negotiating. 72% of a 22.25-point rise is table design. So when a shipper walks into a renewal and asks for a fuel surcharge waiver, it is negotiating a variable the carrier can give away without giving anything up: three points off a table that was drawn by hand can be restored by redrawing it, and the redraw costs nothing but a memo. Waivers expire. Tables persist. What you can actually move sits somewhere else, in three places.

The base rate is one, because it is the only number in the contract that shows up in next year's bid as a comparison. The classification of your ship-to geography is the second, and 2,533 ZIP codes moving on 4 January is the largest single reclassification event in front of you, which means it is also the largest single opportunity to re-price before the market does it for you. The accessorial list is the third: residential, remote, delivery-area, oversize, correction, address-change. Individually small, collectively the same money as the surcharge, and far easier to cap in writing than a formula is.

And the second half of the missed point: 2,533 ZIP codes is a mapping exercise, not a news event. Any shipper with a 2026 ship-to file and a spreadsheet can run the overlay this week. It takes an afternoon. Yet most shippers will find out what it cost them in late January, when the invoice arrives and the budget is already signed, and by then the conversation is about paying the bill rather than about the price. In the end it is a road problem, and the road here is a lookup table. Grace Yang, who always reads these numbers from the buyer's landed-cost side, puts it more sharply than I would: a surcharge column is not a cost, it is a price wearing a cost's clothes, and the moment you book it as a cost you stop asking who drew it.

Now the counterfactuals, because a conclusion you cannot break is not worth much. Suppose diesel goes the other way and climbs hard. Does the surcharge fall? Only if the table is a genuine pass-through, and the index says it is not: three-quarters of the climb came from drawing, and a redrawn table sets a floor as well as a slope. A spike in diesel would add on top of a table already sitting at 23.75%, not travel back down through it. Suppose instead that volumes collapse in 2027 and carriers are hungry. The base rate would soften, and it would soften visibly, because that is the number bids compare. The surcharge table is drawn once and defended after; tables come down later and less than the rates around them. That asymmetry is the whole point.

The disagreement to expect from the other side of the table is a real one and worth taking seriously. A carrier will tell you the surcharge line now carries cost inflation that used to sit in the base rate: wages, insurance, facility, equipment, the cost of serving the thin tail that the network cannot price honestly any other way. There is truth in that. The answer is not to argue about it. The answer is to make the carrier prove it in the structure: if the money is really a cost, then a table with published inputs, a published floor, and a published reset frequency is acceptable to both sides, and a carrier that refuses all three is telling you the column is a price.

What flips my conclusion? Two conditions. If your contract ties the base rate to a published freight index and freezes the accessorial columns for the term, then the surcharge argument stops mattering for you entirely and you should spend your energy on the January ZIP map instead. And if more than about a quarter of your parcel volume sits in the reclassified codes, the right answer stops being negotiation and starts being network design: consolidate, zone-skip, or move the fulfilment node. Negotiation is for the tail you cannot reroute.

Let me put numbers on it, with the assumptions written out. Assume a shipper doing 50,000 residential parcels a year, of which 8% land inside the 2,533 ZIP codes moving to remote on 4 January, so 4,000 parcels. Assume the reclassification adds the published $17.75 a package. That is $71,000 a year of new cost, and it arrives in a single step on a single date. Now take the other side of the ledger and assume the same shipper spends a renewal cycle winning a three-point reduction in the fuel surcharge. Assume an average parcel of $12 in transportation spend, so the surcharge at 23.75% is $2.85 a package and three points is about $0.36. Across 50,000 parcels, that is roughly $18,000. The ZIP map is worth about four times the surcharge fight, takes an afternoon, and does not depend on the carrier agreeing to anything.

Run the LTL side on the same discipline. Assume $1.0m a year of LTL spend priced against the index at 182.5 against a 2018 baseline of 100. A repricing that moves your effective index from 182.5 to 170 is a 6.8% saving, about $68,000, and it comes from freight class, density and packaging rather than from a discount request. The trap in that arithmetic is that a discount request and a re-weigh look identical on a spreadsheet and behave completely differently in year two: the discount is re-bid against you next cycle and the re-weigh is not.

So the work, in order of what it is worth. Run the map first and get a parcel count by 15 November; anything above 3% of your residential base in the affected codes justifies a repricing conversation with every carrier in your mix, not just FedEx. Then get the classification question answered in writing before your agreement resets: ask each carrier to confirm which of your ship-to codes sit in its remote and extended definitions today, and refuse contract language that lets a reclassification pass through as a surcharge change rather than as a rate change. After that, take the fuel surcharge off the negotiating table and put the accessorial list on it; ask for a published table, a floor, and a reset frequency, and if you get all three you have won something that survives the renewal.

Re-weigh and re-class your top twenty LTL lanes before 30 November too, because at 82.5% above the 2018 baseline the classification is worth more than the rate. And price one alternative node before the reclassification lands: zone-skip into a regional carrier's hub, or shift the affected codes to pickup points, and write down the transit days it costs you so the decision is made on facts rather than on a feeling.

The alternatives each have a bill attached, and the bill is always in days or in capital. Zone-skipping to a regional carrier saves the remote charge on the tail and costs you a day of transit and a consolidation step on the front. Pickup points convert a residential stop into a commercial one, which is the cleanest structural answer there is, and costs you conversion: some share of customers will not walk to a locker, and that share is higher in exactly the rural ZIPs where the fee bites. Moving the fulfilment node east or into a denser metro changes the whole zone picture, and costs a lease, a labour market, and six to twelve months. Intermodal on the long truckload lanes takes linehaul dollars off and adds dray, a ramp crossing, and transit days, and the ramp is a node that can clog just as badly as a ZIP table can be redrawn.

Two traps I would point at before you start. One trap is that remote and extended definitions are not the same thing across carriers, and the same ZIP can be remote on one invoice and normal on another; compare the definitions, not the names. The other trap is that minimum charges make a small-volume ZIP look cheap to fix and expensive to fix at the same time: consolidating forty parcels a week out of a remote ZIP into one commercial drop can trip a different minimum and give the saving straight back.

Is this road still walkable? Yes, and that is not a throwaway line. The modes all still work, the parcels still arrive, the trucks still run. What changed is that one of the tolls on this road is no longer a toll; it is a price, and it has been set by somebody who redraws the map once a year and publishes the redraw four days after the holiday. The road is fine. You just have to stop paying at the booth and start reading the map, and the map this year has 2,533 new borders on it, all of them effective 4 January, all of them knowable today.

↑ Back to the key points

— By Ömer Kaya