AutoStore kept 71% of every revenue dollar as gross margin — what is left after the cost of the materials — for the third quarter, in the same release that carried its best-ever third-quarter order intake. Eight days earlier, a survey of 250 warehouse buyers found 94% called switching providers too costly to consider. Set the two figures side by side and the cost of leaving stops being a clause nobody reads. It becomes your exit bill, since the seller books that money as profit and you cannot sell it back.

At a glance

  • The shift: A quarter of buyers regret the system they bought and almost none can afford to replace it.
  • The number: 94% of 250 buyers call a switch too costly to consider.
  • The risk: A payback model that assumes you can leave overstates the return.

The margin and the lock

Exotec published the survey on 28 October, with the fieldwork done between 17 and 31 August among buyers in the United States and Canada [3]. AutoStore reported eight days later, on 5 November. Revenue was $192.1 million, up 38.2%, and gross profit was $137.1 million, a gross margin of 71.3% against 73.1% a year earlier [1]. Orders taken in the quarter came to $177.6 million, up 16.5%, and the company raised its full-year revenue forecast to about $750 million from about $700 million [1]. The same day it agreed to buy Logiqs, a Dutch maker of pallet-handling robots used at about 30 sites, for up to €30 million [2].

The two documents answer different questions and never cite each other. Exotec asked buyers how the choice turned out. AutoStore reported what the choice earns the seller.

I read them together, and the cost line says where the lock comes from. $192.1 million of revenue minus $137.1 million of gross profit leaves $55.0 million of cost of materials, so about 29 cents of each dollar paid for the gear covers the goods [1]. The rest pays for wiring the system into your operation, the software, the start-up, support and the vendor’s profit. None of that comes back out of the building.

That is why the survey’s figure is so high. Sixty-five percent of buyers said total cost of ownership ran above what they expected, and 70% said costs they were not told about raised their yearly upkeep [3]. Sixty percent had a system built by more than one company, and one in seven said they have seen no gain because no vendor owns the result [3]. A system put together that way has no single party to hand back, and that is what makes a swap “prohibitively expensive” for 94% of them [3].

By the numbers

  • $177.6 million of orders: AutoStore’s third-quarter intake, up 16.5% and its best for any third quarter [1].
  • 71.3% gross margin: What is left of each revenue dollar after the cost of materials, down 1.8 points in a year [1].
  • €30 million for Logiqs: What AutoStore paid for a business used at about 30 sites, with 60 staff [2].

Where the lock gets tested

The first test is a number, not an argument. AutoStore’s fourth quarter is guided to lift group revenue to about $750 million for the year [1]. Symbotic reports on 23 November, which puts a second set of order figures on the table before December [4]. What I find hard to square is that both firms can grow while their buyers say they cannot leave. Growth counts new deals. The survey counts old ones.

The second test is policy. Proclamation 11077, signed at the end of October, accepted the finding that imported robots and industrial machinery threaten national security, set no duty, and ordered talks with a report due in 180 days, by 26 April 2027 [5]. Tariffs, quotas and fees stay on the table after that [5]. A ten-year system now carries a landed cost that can move inside its own life, on a clock set by a round the buyer has no part in.

The third test is how buyers are already writing the exit into the deal. Otto Group will put ten AI-controlled dual-arm robots on returns sorting at a Polish site, add 15 more only if the first phase works, and let the vendor run and fix them under a robotics-as-a-service contract [6]. Staging the spend is the plain form of the question here: not what the payback is, but what it costs to stop. The same question sits behind who holds the asset and who holds the price, which is why renting keeps turning up in a market that sells capital gear.

The part neither document can settle

The survey is Exotec’s own, run among 250 people, and its three reasons for regret line up with Exotec’s pitch that one partner should own the whole system. The 94% is a belief about cost, not a measured cost, and the release prints no cross-tab, so nobody can say how many of the regretful buyers are also the ones who cannot leave. I can only set the two shares beside each other and note that they point one way.

The margin is no cleaner. What is left after the cost of materials is not a measure of what the gear cost to build, and it includes software and service that do carry value for the buyer. A fleet that runs for a decade and pays back is still a good buy, even if leaving it would hurt. What I would strike from the model is a resale value at year five, a better seller waiting then, and an exit that costs nothing. Price the cost of stopping first. If that number kills the case, the case was resting on an option nobody bought.

Related reading

Sources