What a Data Center Has to Do With Your Sofa (Part Two)

What a Data Center Has to Do With Your Sofa (Part Two)

Less freight is moving than a year ago and it costs a third more to move it. Nobody did that to furniture retailers on purpose, which is exactly why it will not stop.

By a merchandising executive at a regional U.S. furniture retailer • October 2026 • Submitted anonymously

Missed Part One? Read it here.

In the first half of this I made a case against the ocean carriers. I stand behind it. But I owe you the disclosure that while I was busy building that case, the larger number was moving somewhere else, and I was not watching it closely enough.

Here is the thing that should stop any furniture executive cold.

Dry van volumes in this country are down from a year ago, according to DAT's Truckload Volume Index. Flatbed volumes are down more. Reefer volumes are down hardest of all. By every measure, less freight is moving across America than was moving last summer.

And DAT's spot linehaul rates are up roughly a third year over year. Every equipment type. At once.

Falling volumes and sharply rising prices are not supposed to coexist. When they do, it means the scarce thing is no longer the freight. It is the truck. On DAT's load board, load postings are up from last year while truck postings are down more than 20%, and DAT says contract dry van rates just posted their largest June-to-July gain on record. As DAT’s Dean Croke put it, when rates climb this fast while volumes fall, capacity is what is setting the price. Load postings and actual freight volume are not the same thing, but the distinction makes the imbalance more interesting, not less: even in a soft-volume environment, available truck capacity has fallen faster than demand.

So where did the trucks go?

Three places. None of them are coming back quickly, and only one of them is the story you have probably heard.

Subscribe to Furniture Industry News

One: the purge

The first reason is the one almost nobody in our industry has priced in, because it did not arrive as a freight story. It arrived as an immigration and enforcement story, and most of us filed it under politics rather than under cost of goods.

Federal regulators reinstated aggressive enforcement of English language proficiency requirements for commercial drivers. More than 27,000 drivers have been placed out of service under it, according to a July progress report from the Trucking Association Executives Council that tallied a year of federal enforcement.

That is the small number.

The same report estimates that roughly 194,000 existing non-domiciled commercial license holders (about 97% of that population) will lose eligibility under the Federal Motor Carrier Safety Administration’s new rule. Schneider CEO Jim Filter told investors on the company’s July 30 earnings call that about a third of the non-domiciled drivers he expects to be removed are still on the road. Nearly 10,000 noncompliant driver training providers were struck from the federal registry and 550 fraudulent schools shut down, per the same tally, which means the pipeline that would normally replace those drivers has been narrowed at the same moment. Add 3,200 visa revocations tied to cabotage enforcement, 430 carriers that voluntarily closed their doors, and another 60 to 70 shut down by regulators.

Werner CEO Derek Leathers told investors in July that carrier exits are “still in the early innings.”

Whatever you think of the policy (and I am not writing to argue about the policy), the freight consequence is not in dispute. A very large number of drivers are being removed from the American road on a schedule set in Washington, and the schools that would have backfilled them are being closed on the same schedule. That capacity is not sitting on the sidelines waiting for rates to improve. It is gone.

Two: the bleed

The second reason is the one every trucking executive has been describing for four years, and it is real even though we have all gotten bored of hearing it.

IntelliTrans’ Blake Azell estimated on FreightWaves that the industry has lost roughly a quarter of a million drivers since 2020. On that same July call, Schneider’s Filter said the long-haul driver population sits well below its long-term average and near the lowest level of the past decade.

This is a demographic problem wearing a labor-market costume. The people leaving are older than the people entering, there are more of them, and the entry path just got narrower.

On its own, that produces the slow grind the industry has lived with since the pandemic. Stacked on top of a regulatory purge, it produces something else. Schneider’s Filter said the quiet part plainly on that call: capacity will not grow back the way it did after the pandemic, and that is precisely why this cycle can run longer than the last one.

Read that as a furniture retailer rather than as a trucking analyst. He is not forecasting a spike. He is forecasting a floor.

Three: the bidding war we did not know we entered

I had heard the third reason the way you probably have: as a rumor that the AI build-out is eating the trucking market. I went looking for the arithmetic half expecting it to fall apart.

It does not fall apart. It is worse than the rumor.

A single 500-megawatt data center takes an estimated 30,000 truckloads of concrete, steel, copper, fiber optics and generators, according to figures cited by FreightWaves. Not across its operating life. To build the thing. The hyperscalers have announced capital programs in the hundreds of billions of dollars. FreightWaves put the pace this way: they are spending every two weeks what the interstate highway system cost per year to build.

And here is the number that should ruin your afternoon: FreightWaves counts only about 8% of announced and contracted data center projects as under any level of construction.

This is not the peak. This is the warm-up.

IntelliTrans’ Blake Azell has a name for what that does to a freight market. He calls it a capacity tax: the premium everybody else pays because one buyer who does not care about the price has entered the room. Put the two of us side by side at the same loading dock, bidding on the same truck.

When I pay more for that truck, the money comes out of the margin on a sofa. There is a ceiling on what I can pay, and I know exactly where it is, because past a certain rate the SKU stops working and I stop buying it.

When they pay more for that truck, the money comes out of a construction schedule. Every week a finished facility sits waiting on a late shipment of switchgear is a week that an enormous capital investment earns nothing at all. Against that, the freight premium is a rounding error, and the later the project runs, the more they will pay to keep it moving.

We are not bidding with the same money. The truck goes to whoever has the larger number sitting behind it, and it is not the sofa.

“But furniture moves in dry vans”

Here is the objection, and it is a fair one. Data centers move on flatbeds and heavy haul. Furniture moves in dry vans. Different markets. Nobody is loading a bedroom set onto a step-deck bound for Abilene.

The objection does not save us, for three reasons.

The first is substitution. Flatbed spot rates now run well above dry van on DAT’s load board, with dozens of loads chasing every available trailer, and equipment and drivers move toward the money. A carrier weighing a premium construction lane against your furniture lane is not being disloyal when it takes the construction lane. It is being solvent. Every truck that converts is one that left your market, and you never see it happen. You just see the rate.

The second is that the driver shortage does not respect equipment type. A driver removed from the road by an out-of-service order does not come back as a dry van driver. He does not come back at all. The purge and the bleed subtract from the same pool that hauls your sectionals.

The third is the one furniture retailers should be most alarmed by, because it hits the part of the operation we actually control.

We run our own home delivery. We hire our own drivers and our own warehouse labor, out of our own local markets. Data center construction pays a substantial premium over ordinary construction wages, and the skilled trades are already short-handed nationally. In the towns where these campuses land (Abilene is one, and Texas has more coming), that premium resets the wage floor for anybody in the county who can drive a truck or move heavy freight.

So we pay for the AI build-out twice. Once in the linehaul rate, through capacity lost to a bidder we cannot beat. And once in our own delivery payroll, competing for the same people in the same towns.

Why this is a strategy problem and not a freight problem

Everything in the first half of this series was a fight I could have. There was a counterparty, a contract, a tariff on file, a competitor who wanted my business. I could pick up a phone.

There is no phone call here.

Nobody at a hyperscaler is thinking about furniture retailers. No regulator writing a licensing rule is weighing what it does to the cost of delivering a sofa in the heartland. No driver taking a better-paying construction lane owes me an apology, and I would take that job too. There is no villain in this half of the story, which is exactly what makes it worse, because a villain can be negotiated with and a structural shift cannot.

On the water, somebody is doing this to me on purpose. On the highway, nobody is doing it to me at all. That is the more expensive of the two.

Which means the reflex our industry built over the last four years (treat freight as a temporary emergency, absorb it, wait for normal) is now the most costly habit in this business. There may be seasonal relief, but retailers should no longer assume seasonal relief will restore the freight economics of the pre-tightening market. There is no quarter where 194,000 licenses quietly come back.

This is a structural reset in the cost assumptions furniture retailers should be using for the next several years, arriving on a timeline set by people who have never heard of us.

What to actually do about it

You cannot negotiate with this one, which changes what you do rather than whether you do anything. Every item below is something a merchandising organization controls without asking a carrier for permission.

Rebuild the freight assumption underneath your assortment. Model the next three years on structurally higher truckload rates and a structurally tighter driver market, and then look at what breaks. If your delivery radius, DC footprint, free-delivery threshold and low-price-point assortment were designed when freight was cheap, they were designed for a country that no longer exists. The bulky, low-margin SKU that barely worked at old freight rates does not work now, and the sooner that shows up in a line review the better.

Put cube economics back in front of buyers. When freight is cheap, cubic efficiency is a rounding error and nobody argues about it in a buy meeting. At current rates it is margin. Knock-down construction, tighter packing, container fill rate, and how many pieces actually fit in a forty-foot high-cube deserve to be selection criteria and not afterthoughts.

Find out what is being built within a hundred miles of every facility you run. Data center construction is public information long before it breaks ground. If a large campus is coming to one of your markets, your delivery wage problem has a start date. You would rather know that now than discover it when a third of your drivers leave in a month.

Treat driver retention as a merchandising issue, not an HR issue. If the local wage floor is about to move, the cheapest driver you will ever employ is the one you already have. Every experienced delivery driver who leaves takes route knowledge, damage rates and customer experience with him, and you will replace him at the new market price whether you planned for it or not.

Get honest about who is really paying for free delivery. A delivery promise written against old freight economics is a discount you are funding out of gross margin without saying so out loud. Either price it, restructure it, or defend it deliberately, but stop letting it drift.

Quote landed cost as a band, not a number. Between ocean surcharges you do not control and a domestic leg on a multi-year escalator, a point-estimate landed cost is fiction, and you are setting retail prices on top of it. Price the range. Hold reserve against it.

I spent the first half of this series arguing that the ocean carriers have built a machine for holding prices the market has already rejected. I believe that, and I think our industry should push back on it much harder than it has.

But if I am honest about where the bigger money is going, it is not there.

It is in a domestic freight market that is quietly repricing itself around a driver population being shrunk by policy, a workforce aging out faster than it is replaced, and a construction cycle so large it has made trucking capacity into something furniture retailers now bid for against companies that do not notice we are in the room.

The ocean fight is worth having. It is just not the one that decides whether the delivered cost of a sofa in the middle of the country works three years from now.

Read every invoice, not the summary. Ask for the formula. And then go find out what is being built down the road from your warehouse.

---

If someone in our industry would get something out of this, please forward it to them.

---

Get Furniture Industry News in your inbox, free: https://www.furnitureindustrynews.com/subscribe/

Subscribe to Furniture Industry News

---

Powered by beehiiv