AutoStore Cube Storage System
Photo: AutoStore
- energy
- 10 robots consume roughly as much as one household vacuum cleaner
- spec note
- This pass re-verified volatile fields only. No new robot generation was launched and nothing in the current lineup was discontinued.
- current robot
- R5 Pro (launched 2023, no successor announced)
- bin heights mm
- 220 / 330 / 425
- uptime percent
- 99.6
- bin stack depth
- 26
- avg bins per system
- 34000
- go live weeks midsize
- 26
- competing solution lead time weeks
- 52-78
Who's exposed
Jobs in the threat radius
- warehouse picker
- storage zone operator
- inventory clerk
- putaway worker
- order fulfillment associate
Deployment status
Re-verified, no material change — which is itself worth recording honestly. Every scale and financial figure was checked line by line and confirmed correct. As of Q1 2026: more than 1,950 active systems worldwide, over 1,300 customers, roughly 2,000 sites, 65-plus countries. Q1 2026 revenue was USD 165.8M (+92.9% year over year, -7.8% quarter over quarter), order intake USD 179.4M (+27.0%), backlog USD 570.6M (up from USD 557.0M in Q4 2025), gross margin 72.7%, adjusted EBITDA margin 44.0%, cash conversion 81.9%. Cross-check: the March 2026 product release cites '1,900-plus systems in over 65 countries' — consistent and earlier. AutoStore discloses no robot count and no port count in any 2026 filing. Mats Hovland Vikse remains CEO, unchanged. No restructuring and no layoffs; the Q1 report in fact records higher employee benefit expense and continued investment in long-term growth initiatives. No new robot generation exists: the most recent launch was the Spring 2026 portfolio on 2026-03-19 (CubeVerse platform, AutoStore Intelligence, CubeStudio, CubeAnalytics, CubeControl, VersaAI), all software and AI, and all predating this entry's last verification. The R5 Pro remains the current robot, launched in 2023 with no announced successor, and Pio remains active for the SMB segment. One open question: CarouselAI, an award winner in 2025, is absent from the March 2026 launch while the newly introduced VersaAI covers the vision and piece-picking function. That may be a rename or a supersession, but no announcement confirms either, so we do not record a discontinuation. Litigation: all patent claims with Ocado were fully settled on 2023-07-24, with AutoStore paying GBP 200M (about USD 256M) in instalments, all global claims withdrawn and a cross-licence on pre-2020 patents. No new litigation was found. One timeliness warning that belongs in the record: AutoStore's Q2 2026 results are scheduled for 2026-08-13, four days after this re-check. Every financial figure above will be superseded then, and this entry should be re-checked on or after that date.
When this hits the labor market
Unchanged; nothing in this re-check justifies a revision. Already deployed at scale. Over 1-3 years, the next wave of large retailers and 3PLs that have not built yet. Over 3-5 years, the long tail of mid-sized operators, as Pio lowers the floor for the SMB segment. The one observation worth adding comes from the pricing structure rather than the technology. AutoStore's pay-per-pick model compresses the upfront grid infrastructure payment to 20-40% of total cost and converts the rest into a subscription that scales with order volume. That has nothing to do with robot performance and everything to do with how the system gets compared against people: once the billing unit changes from a one-time multi-million-dollar capital expenditure into a price per pick, it lands on the same page as a picker's hourly wage. That is this category's real diffusion mechanism, and it is also why a 26-week go-live against competitors' 52-78 weeks determines who gets displaced first far more than any hardware specification does.
Is this robot coming for your job?
Press releases speak in averages — LostJobs talks about your role. Find out which parts of your job hold up and which don't. Free.
Talk to LostJobs about my future
Why “no change” gets written down
Because this catalogue shows readers a “Verified by LostJobs.AI” date, and that date only means something if somebody actually went and checked. A record nobody has looked at in three months and a record that was just walked through line by line and confirmed correct look identical on the page. The only difference is that one of them is true.
So, for the record: this pass checked AutoStore’s status, pricing structure, scale figures, financials, product line, leadership and litigation posture. All confirmed. Nothing required amendment.
A timeliness warning — revisit this in four days
AutoStore’s Q2 2026 results are scheduled for 13 August 2026, four days after this re-check.
Which means every financial and scale figure in this entry — 1,950 active systems, USD 165.8M of quarterly revenue, USD 570.6M of backlog — will be superseded within days. That is not a failure of this pass; it is what happens when an earnings calendar and a verification calendar fall out of phase. But readers should know, and the next cycle should put this entry at the front of the queue.
What was checked
Scale. More than 1,950 active systems, over 1,300 customers, roughly 2,000 sites, 65-plus countries as of Q1 2026. The March 2026 product release’s “1,900-plus systems in over 65 countries” is earlier and consistent, which cross-checks. AutoStore discloses no robot count and no port count anywhere in its 2026 filings.
Financials. Q1 2026 revenue USD 165.8M (+92.9% YoY, -7.8% QoQ), order intake USD 179.4M (+27.0%), backlog USD 570.6M, gross margin 72.7%, adjusted EBITDA margin 44.0%.
Product line. No new robot generation. The R5 Pro remains current, launched 2023, no successor. The Spring 2026 portfolio of 19 March 2026 — CubeVerse, AutoStore Intelligence, CubeStudio, CubeAnalytics, CubeControl, VersaAI — is entirely software and AI, and predates the last verification. Pio remains active for SMBs. One open question: CarouselAI, which won awards in 2025, is absent from the March 2026 launch while VersaAI now covers vision and piece-picking. Possibly a rename, possibly a supersession; no announcement confirms either, so we record no discontinuation.
Leadership and litigation. Mats Hovland Vikse remains CEO. No restructuring, no layoffs — the Q1 report shows rising employee benefit expense. The Ocado patent dispute was fully settled on 24 July 2023, with AutoStore paying GBP 200M (about USD 256M) in instalments, all global claims withdrawn and a cross-licence on pre-2020 patents. No new litigation found.
A methodological note: AutoStore’s news index renders without dates. We date-verified each item before treating anything as 2026 news, and two items that looked current turned out to be from 2025 and 2023. That page structure makes it very easy to publish old news as new.
Whose work it takes
Warehouse pickers, storage zone operators, inventory clerks, putaway workers, order fulfillment associates — unchanged, as is the customer base: Walmart, Best Buy, Puma, Lululemon, John Lewis, ASDA, ABB and hundreds of others.
The mechanism is worth restating because it differs from the mobile robots elsewhere in this catalogue. AutoStore does not bring the shelf to the human. It abolishes the aisle. Bins stack 26 deep, robots run on a grid across the top, and people do not enter the storage volume at all. This is not a picker walking less. It is nobody being in there. The storage zone operator role disappears wholesale at deployed sites.
The 99.6% uptime figure and the “10 robots draw about as much as a household vacuum” line are two ways of saying the same thing: almost nothing in this system’s operating cost structure scales with headcount.
Why we care for LostJobs
Because the most useful finding in this re-check is not in the technology column. It is in the pricing column.
AutoStore’s pay-per-pick model compresses the upfront grid payment to 20-40% of total cost and turns the remainder into a subscription that floats with order volume. That has nothing to do with how good the robots are, and everything to do with how the system gets compared to people. Once the billing unit stops being a multi-million-dollar capital expenditure and starts being a price per pick, it sits on the same sheet as a picker’s hourly wage.
Capital expenditure has to persuade a board. A variable cost only has to persuade a spreadsheet.
Worth remembering alongside it: 26 weeks to go live, against 52 to 78 weeks for competing solutions. In this category, who gets displaced first is decided less often by which machine is better than by which system can be installed inside the current fiscal year.