Supply Chain Action Points
Read this first — the conclusion, and the moves to make:
- By 30 June, tender three to five representative lanes to the new entrant, including at least one consignee with no dock and one with a tight receiving window, and measure on-time delivery and damage for 30 days before shifting volume.
- Get written answers on who physically executes delivery outside the dock-to-dock footprint, what liftgate and appointment lead times apply, and what the re-delivery and detention charges are per incident.
- Cap single-source exposure on every lane at 60% through the fourth quarter, with a named secondary carrier on each lane and a documented switch trigger.
- Add a 30-day service-withdrawal notice clause and an accessorial cap to the routing guide before the Q4 peak, and require notice of any change to appointment scheduling lead time.
- Re-bid the incumbents on total landed cost per lane including accessorials, not on base rate, and target a 3% reduction in total cost rather than a 3% cut in linehaul.
Amazon Freight officially opened its less-than-truckload service to all shipper segments and destination types, ending a long-standing pilot that limited access to mid-volume and large national accounts. The expansion leverages Amazon's existing middle-mile linehaul network and a growing fleet of 53-foot dry van trailers, with delivery promises comparable to national LTL carriers on key lanes. The move comes just weeks after FedEx Freight's NYSE debut as FDXF and ArcBest's mid-cycle GRI, suggesting Amazon is positioning to capture more of the post-Yellow LTL wallet share. Traditional carriers are expected to respond with denser service maps, accelerated transit guarantees, and tiered pricing for tech-forward shippers.
The Analysis
An LTL network is a dock network with trucks bolted to it. The trucks are the visible part and the least interesting one. Amazon Freight has opened its less-than-truckload service to all shipper segments and all destination types, ending a pilot that kept it to mid-volume accounts and large national ones, and the question worth asking is not whether Amazon can sell LTL. It is which dock the freight will actually sit on.
The assets behind the announcement are specific. Amazon is using its existing middle-mile linehaul and a growing fleet of 53-foot dry vans, and it is promising delivery comparable to national LTL carriers on key lanes. That is a real offer on lanes that fit the shape of its network. It is not a national LTL network in the sense ArcBest or Old Dominion runs one, and the gap between those two things is where a shipper gets surprised in the fourth quarter.
The timing tells you what this is. The announcement lands weeks after FedEx Freight began trading on the NYSE as FDXF and weeks after ArcBest's mid-cycle general rate increase. Yellow's exit left a large block of national LTL volume looking for a home, and every carrier in the country has been circling it. Amazon is not walking into a calm market. It is walking into an auction.
Count the touches on one LTL shipment and the service promise stops sounding abstract. The freight is picked up on a local route, taken to an origin terminal, unloaded, sorted, reloaded onto a linehaul trailer, driven overnight, broken apart at a second terminal, reloaded onto a delivery route, driven to the consignee, and unloaded at the door. Nine touches, nine chances to put a fork through a carton or lose a pallet label. The linehaul between terminals is the easiest part of that chain and it is the part Amazon already owns.
The capability that runs short is at the two ends, and it is not tractors. It is doors and appointment slots. Amazon has doors at its own buildings, where it controls the schedule, the staff and the yard. Your consignee has one door, a receiving clerk who goes to lunch at noon, and a booking system that is already full on Thursday. A 53-foot dry van is a dock-to-dock tool. It needs a raised concrete platform and a forklift. The announcement says all destination types, and that is a commercial statement, not an equipment statement. A strip mall with a curb and no ramp does not become dock-height because a rate sheet says so.
The second constraint is the shape of the network, and this is the part I would want a shipper to sit with. Middle-mile linehaul is bidirectional and it is unbalanced. Amazon moves heavily into its fulfilment centers and lighter on the way out, which is exactly why selling LTL to outside shippers makes financial sense: it monetises the weak direction. That is genuine capacity, and it is priced attractively for a reason. It is also conditional capacity. It exists because Amazon's own freight is not using that leg on that day. In October, November and December, Amazon's own freight will use it.
The causal chain from Yellow to this announcement is short and physical. A large national LTL carrier stopped operating, which removed a block of dock capacity, tractors and drivers from the market in one weekend. FedEx Freight separated and listed, which means it now answers to public shareholders every quarter. ArcBest moved its general rate increase into the middle of the cycle rather than waiting for the usual window. Into that, Amazon brings a national promise built on a network that was designed for its own parcels and its own buildings. Traditional carriers are expected to answer with denser service maps, faster transit guarantees and tiered pricing for technology-forward shippers, and every one of those answers costs money that has to come from somewhere in the invoice.
Who gains depends on what they could buy before. A small shipper that could never open a national LTL account now has a national network at a published price, and that is a genuine change in its options, not marketing. A mid-size shipper gets a third bid on lanes where it previously had two. A large shipper with dense lane volume will get better attention from the incumbents, and should expect that attention to arrive as accessorial changes rather than base rate cuts, because base rates are the number everybody compares and the accessorials are where the margin lives. Exporters should be careful here. This is a domestic middle-mile and distribution product. A plant-to-port move with a vessel cut-off is a different animal with different equipment and a hard deadline, and a network built on fulfilment centres will not save you when the sailing is at midnight.
The timing runs on three clocks. Onboarding runs on days: the first month is account setup, EDI mapping and lane profiling, and the service numbers are noise until that settles. Price runs on one to two quarters, because that is how long a routing guide takes to re-tender. Service runs on Amazon's own calendar, and the honest test is the fourth quarter, when its own volume comes back and third-party freight is competing for the same trailers.
Here is the point the coverage does not make. The thing being sold is not spare capacity in the way a new terminal is spare capacity. It is the unused half of somebody else's loop, and the seller controls the loop. That does not make it a bad product, it makes it a product with a season attached. Ask what happens to your tender in week forty-eight, and ask it in writing, before you build a routing guide around it. The second half of that point is about equipment. The 53-foot dry van fleet defines the real serviceable footprint, and everything outside that footprint is executed by somebody else. When a shipment needs a liftgate, a residential delivery or inside placement, you are not buying Amazon's network any more. You are buying a subcontractor whose name is not on the rate sheet, and that handoff is where the damage claims, the re-delivery fees and the missed appointments actually live.
A calculation, and the numbers are my assumptions rather than anything in the announcement. Take ten lanes and move two of them to the new entrant. Suppose those two lanes carry a hundred loads a month. If five loads miss the delivery appointment and each miss costs a two-hundred-and-fifty-dollar redelivery charge plus a hundred and fifty dollars of dwell, that is two thousand dollars a month of leakage, which is roughly two percent of the spend on those lanes if the average invoice is around a thousand dollars. State the assumptions plainly and the arithmetic is not the point. The point is that a two percent rate advantage disappears the moment the appointment discipline is worse than the incumbent's, and appointment discipline is exactly what a dock-to-dock network is worst at when the destination is not its own building.
The counterfactual matters for anyone drafting a bid this month. If Amazon had stayed inside the pilot, your routing guide would have kept two carriers per lane and you would have paid the mid-cycle increase without an alternative. The alternative is real. It is also narrower than the headline, and a shipper who treats one national promise as full coverage will find out in December which half of the network it actually bought.
The trap is comparing base rate to base rate. Two carriers can quote the same linehaul number and invoice completely different totals once the liftgate, the appointment fee, the re-delivery and the detention are counted. Put three lanes out to tender this week, deliberately including one consignee with no dock and one with a tight receiving window, and measure on-time and damage for thirty days before you move volume. That is a cheap experiment, and it tells you more than any press release will.
↑ Back to the key points