Amazon Opens Full-Scale Less-Than-Truckload Freight Network to All Businesses

The Analysis

Yesterday a customer was still on the phone with me, halfway through explaining why a pallet of goods had to be on a retail dock by Thursday noon or the whole promotion would stall. That is the kind of call I get all day on the last-mile desk. When freight moves on a less-than-truckload network, the gap between a clean delivery and a missed appointment is almost never the number on the rate sheet. It is the hour the dock cuts off, the hour the truck actually shows, and whether the handoff between the line-haul and the local terminal happened before somebody went home for the night.

So when Amazon Supply Chain Services opened its full-scale less-than-truckload network to all businesses on June 10, 2026, my first reaction was not that cheaper freight had arrived. It was a question about the clock. Amazon had run an inbound-only LTL model for its own sellers, and now it is selling pallet space on that same network to anyone. Sharing trailer space for partial loads is how you cut cost against a full truckload. A rate everyone can see is not the same thing as a delivery everyone can trust at the hour it was promised.

The story the trade press is telling is a market-share story. Amazon is now a direct competitor to FedEx Freight, now trading as FDXF, and to XPO and Old Dominion, leaning on infrastructure and a technology platform it already owns. The market it is walking into is already tightening because FMCSA enforcement has been pulling capacity out of the lane. New capacity sounds like relief. From where I sit, the relief is real but it is not free, and the part that costs you is the part nobody measures on day one: how steady the delivery actually is.

Let me slice the facts before anyone gets excited. On June 10, 2026, Amazon Supply Chain Services launched a full-scale LTL freight offering open to all businesses. Before this it was inbound-only, built for Amazon sellers moving goods into Amazon's own network. Now any business can ship by pallet, sharing trailer space with other partial loads instead of paying for a whole truck. The stated competitors are FedEx Freight, now rebranded FDXF, plus XPO and Old Dominion. And the market context the announcement rides on is a lane that has been tightening because FMCSA enforcement has been removing capacity. Those are the only hard facts we have. Everything else in the headlines is interpretation, and interpretation is where shippers lose money if they are not careful.

Why now and why this shape. Amazon did not build a trucking company from zero. It already ran pallet freight for its sellers on the inbound side, so the trailers, the terminals, the scan system, and the driver routing were already turning. What changed on June 10 is that the spare seats on those trailers became a product anyone can buy. That is the cheap part of the move: monetize space that was already rolling. The push from the market side is FMCSA enforcement thinning incumbent capacity, which opens a gap a new seller of trailer space can step into. So the cause is not a sudden drop in freight rates. It is spare capacity meeting a tightening lane, and a tightening lane is exactly where a new seat to buy looks attractive to a shipper who cannot find one.

Who actually feels this first. The shippers who win are the ones moving non-urgent pallets, the stuff where a delivery landing Tuesday instead of Monday does not break a promise to anyone downstream. A distribution center replenishing slow-moving stock, a manufacturer feeding a warehouse that holds buffer inventory, a retailer backfilling shelf stock between promotions: those are the accounts that can ride a new network while it finds its rhythm. The shippers who should be nervous are the ones carrying a time commitment on every pallet, the promotional display that must be in store by a fixed morning, the contract with a penalty clause if the dock appointment is missed, the assembly line that stops if the component pallet is late. For them a new network is not a discount. It is a variable they do not yet control, and a variable you cannot control is the one that shows up in a complaint, not in a savings line.

What the capacity tightening really means on the desk. FMCSA enforcement has been pulling capacity out of the lane, which is the reason a new seller of trailer space finds an opening. But here is the trap I watch for: when incumbents get thin, their own cut-off discipline can slip too, because a carrier running hot is a carrier that misses its own dock-close time. So the new capacity is most useful precisely in the weeks incumbents are full, and those are the same weeks a shipper is tempted to throw promise-bound freight at any open trailer. Useful and dangerous arrive together. The new network gives you a seat. It does not give you a promise you can bank yet, and the difference between a seat and a promise is the whole point of this piece.

How to read a published cut-off before you trust it. A new entrant posts a dock-close time, and on paper it looks like every other carrier's. The number I want is not the posted time. It is the gap between the posted time and the time the last trailer actually leaves, measured over a few weeks. A dock that says 5 p.m. but regularly rolls the last load at 6:30 is a dock that will quietly eat your 5 p.m. promise, and you only learn that by watching, not by reading. This is the kind of variance a brochure never shows and an incumbent has already worn smooth.

Pickup cadence is the next thing that betrays a new network. On an incumbent lane you know whether your zip gets swept at 9 a.m. or 2 p.m., and you build your own cut-off around it. On a new network the sweep may be once a day, or every other day in some zips, and a pallet that misses the morning sweep waits a full day before it moves at all. A full day of wait before the freight even enters the system is a day of variance no rate discount pays for, because the promise clock started the moment the customer expected the pallet, not the moment the truck finally showed.

The line-haul transfer is where I would put my own money on the risk. Amazon's trailers run, but a network tuned for Amazon's own inbound flow picks up at Amazon facilities and drops at Amazon facilities. Outside freight enters and leaves at points that network was not built around, and the break-bulk handoff is the hour a mixed pallet can go quiet. Quiet is the word I use for a pallet nobody is actively moving and nobody has yet owned. On a mature network the handoff is owned by a routine. On a new one it is owned by whoever happens to be standing there, and that is not a promise you can put in a contract.

What service maturity looks like week by week, if you are patient. The first two weeks of a new network are noise: everyone is careful, the launch is watched, the easy freight moves. Weeks three through eight are where the truth shows, because the careful phase ends and the real mix of freight arrives, including the awkward shapes and the tight zips. If the on-time spread narrows across those weeks, the network is learning your lane, and that is the only signal that earns the next slice of volume. If it stays wide, or widens, you have your answer without ever risking the freight that cannot slip. Maturity is not a date on a press release. It is a trend you read on a spreadsheet every Friday.

One more thing the headlines get backwards. They frame this as Amazon versus the incumbent carriers, a fight for share. From the desk it is not a fight you are watching. It is a choice you are making, pallet by pallet, about which network carries the freight that breaks your promise if it is late. The incumbents are not your enemies here. They are the benchmark the new network has to beat on consistency before it touches your urgent freight. Treat the launch as a tryout, not a turnover, and you keep control of the only number that reaches your customer: the hour the pallet lands.

The trap of the headline discount, spelled out for the desk. A few points off the rate sounds like free money when you read the launch story. Free money is rare in freight, and when it appears it is usually priced somewhere else, in a variance you have not measured yet. The discipline that protects you is boring: do not let a discount decide which freight moves. Let the on-time record decide, and let the discount be a bonus you collect only after the record clears your bar. A shipper who flips volume for a rate and discovers the variance later is paying retail for a lesson the spreadsheet would have taught for free.

Why I keep saying hour and not day. A day is a unit a planner uses. An hour is a unit a promise uses. When a retail dock closes at noon, the difference between 11:50 and 12:10 is not twenty minutes. It is a made appointment and a missed one, and a missed appointment is a penalty, a clawback, a relationship dent. The new network's value to you is decided in those ten-minute gaps, not in the quarterly capacity story the press loves. That is the whole reason my knife cuts by the hour, and the reason a launch measured only in market share tells you nothing about whether your Thursday pallet lands.

Now I take the knife I actually use, the one that runs along the shipment clock, and I cut this new network hour by hour. The path is cut-off, pickup, line-haul, the terminal handoff, and delivery. On the cut-off: a new entrant has to publish a dock-close time and then hold it, and holding it is a discipline built over months, not announced in a press release. On pickup: the cadence, how many times a day the truck sweeps a given zip, is what decides whether your pallet waits six hours or sixteen before it even moves. On line-haul: Amazon's trailers are already running, so the long move may be fine, but the question is the transfer between line-haul and the local terminal where the load gets broken and re-stacked.

That is the hour where a new network either proves itself or drops a pallet on the floor of a sortation center at 2 a.m. with nobody accountable yet. For domestic LTL the customs step is light, but the terminal sort is the real border between moving and stuck, and it is exactly the step a network built for Amazon's own inbound flow may not have tuned for outside shippers with mixed freight. Then delivery: the last mile is where the customer's promise lives, and a new last-mile footprint is the slowest thing to mature.

Here is what the headlines did not tell you, and it is the only part that should change your behavior. Everyone wrote that Amazon disrupts LTL and prices drop. The number that matters to a shipper is not the rate. It is the variance. A new network's on-time average can look fine in a brochure while its hour-by-hour spread is wide, meaning some pallets land early and some land a day late, and the late ones are exactly the ones attached to a promise. Saving a few points on the freight rate can be wiped out by one missed dock appointment with a big account, the penalty on the contract, and the phone call where you explain why their promotion did not open on time. The so-called new capacity is not an immediate cure for a shipper who is in a hurry this week. It is capacity with a maturity curve attached, and the curve is the cost nobody put in the headline.

The second thing a brochure hides is claims and returns. When a pallet is damaged or lost on an incumbent network, you know the claims desk, the turnaround, the person who picks up the phone. On a network that opened to all businesses on June 10, that machinery is brand new for outside freight. Damage rate and claims turnaround are the two numbers I would watch before I move a single urgent pallet. A lower rate on the invoice means nothing if a damaged display takes six weeks to get paid back and the promotion is already over. The maturity of the claims process is invisible on day one and expensive on the first claim, which is exactly why you do not learn it the hard way on freight that cannot slip.

When does any of this hit your books. Not next year. From the first pallet you route onto the new network. The variance shows up on shipment one if you put the wrong freight on it. The claims maturity gap shows up on the first damage, which on a new mixed-freight sortation flow tends to come early. The relief from new capacity, if it is real, shows up as available trailer space in the weeks when incumbents are full because FMCSA enforcement thinned them. So the timeline is not a slow burn. It is a choice you make per pallet, and the choice decides the outcome on the very next dispatch, not on some quarterly review.

What would make me wrong. If Amazon had been running this exact pallet product for outside businesses for three years, the variance would already be beaten down and my caution would be overcooked. It has not. It opened June 10. The condition that flips the conclusion is service maturity, and maturity is a function of time on the lane, not size of the parent company. Amazon is huge, but a huge company running a new product for outside shippers is still a new product for outside shippers. Size buys you trailers. It does not buy you the rhythm of a dock that has learned your freight, and rhythm is the only thing a time commitment is made of.

So how do you take the new capacity without eating the variance. Start by segregating the freight. Put only the freight whose lateness costs you nothing onto the new network first: replenishment to a warehouse with buffer, slow-moving stock, anything where a day of slip is invisible downstream. Keep the freight with a fixed delivery promise on the incumbent you already trust until the new network earns it. Before you shift any real percentage, run a parallel lane for a few weeks and watch on-time rate, damage rate, and claims turnaround as the only three numbers that matter. And write the acceptance bar in ink before the first pallet moves, so when the variance shows up you already know whether it passed, instead of arguing about it after a customer is angry.

Let me run the math the way I would in a meeting, with the assumption spelled out so nobody confuses it with a fact. Assume, as an illustration and not as a reported number, that a customer shifts X percent of its LTL volume from its incumbent carrier to Amazon's new network. Assume further that because the new network's delivery variance is still unproven, Y percent of those shifted shipments arrive outside the promised hour window. Assume each late shipment carries a contractual penalty of Z dollars, whether that is a dock appointment fee, a promotion clawback, or a service credit the customer demands back.

Then the expected penalty from the shift is, in plain terms, X percent of volume times Y percent late times Z dollars per late shipment. The freight saving on that same X percent of volume is some smaller number, say a few points off the incumbent rate. The point of the model is not the exact figures. It is the shape: a saving measured in rate points can be overrun by a penalty measured in broken promises, and the broken promise lands on the shipments you least wanted to risk. If your Y times Z is bigger than your rate saving, the discount was a trap. Put your own X, Y, and Z in and the answer is yours; I am not putting numbers on Amazon's network that nobody has published.

On my desk the cut-off hold rate is the tell I trust most, because that is where a new entrant either keeps its word or does not. If the published dock-close time slips by more than a small window in the first month, the pickup cadence is not real yet and I keep urgent freight off. After that I watch the terminal handoff hour, the time the pallet actually leaves the sortation center toward the local driver. If that handoff drifts past the window where a driver is still on shift, the pallet sits, and sitting is how promises die. I would rather see a new network miss by a consistent thirty minutes every time than hit by luck and miss by a day the next. Consistency is the thing a new entrant has to earn, and it is the only thing that makes the rate worth taking, because consistency is what your own customer's promise is built on.

None of this is a reason to refuse the new capacity. It is a reason to treat it like a new driver on a route you care about: prove the rhythm before you hand over the freight that breaks your promise if it is late. The market is tightening because FMCSA enforcement took capacity out, so having another seat to buy is genuinely useful. Just buy the seat for the freight that can wait, and keep the freight that cannot wait on the carrier that has already learned your dock. The clock does not care how big the parent company is. It only cares whether the pallet is on the truck at the hour the promise was made.

↑ Back to the key points

— By Nadia Pryce