On 26 October 2026 Agility Robotics and GXO Logistics announced they are expanding their partnership to two more sites — Statesville in North Carolina and Lebanon in Tennessee. Deployments are due to start in early 2027 on Digit 4 robots. It is the first expansion in the three years since a pilot widely regarded as the first commercial humanoid deployment in a live logistics operation.
Key takeaways
- Two new sites: Statesville in North Carolina and Lebanon in Tennessee, supporting e-commerce fulfilment
- Deployments start in early 2027 on Digit 4, migrating to Digit 5 once it is commercially available
- At Flowery Branch in Georgia, Digit has moved more than 100,000 totes
- Agility reports more than 65,000 hours of real-world robot operation
- In August, Patrick Kelleher conceded that after 45 pilots humanoids have yet to produce a return on investment
Where this started
A warehouse in Flowery Branch, Georgia, late 2023. A humanoid robot gets one job: take a tote off a rack and place it on a conveyor. No acrobatics.
Three years on the counter reads more than 100,000 totes moved. This is not a demo. It is one repetitive task, performed alongside existing automation.
But here is the problem: GXO still has not worked out when that work starts paying for itself.
That is the backdrop against which today's expansion has to be read.
What exactly did the two companies agree?
Agility Robotics will supply Digit to two new GXO locations. At Statesville and Lebanon the robot is to support e-commerce fulfilment, taking on highly repetitive material-handling work.
Deployments start in early 2027 and will run on Digit 4. Migration to Digit 5 follows once that model becomes commercially available.
The robot runs on Agility Arc — a cloud platform for managing robot fleets and integrating them into customer operations.
This is a simplified timeline based on the press release, not a full deployment schedule.
What do these names mean?
A handful of acronyms recur in this announcement and change the meaning of a sentence.
- RaaS (Robots-as-a-Service) — a model in which the customer pays for the robot's work rather than owning the robot. A subscription instead of a hardware purchase.
- 3PL — a third-party logistics operator running a warehouse on another company's behalf. GXO describes itself as the world's largest pure-play contract logistics operator.
- Tote — a standard warehouse bin. The unit in which the robot's work is counted here.
- ROI — return on investment. The point at which savings exceed deployment cost.
- Physical AI — AI driving a physical body rather than only text. Both companies use the term for the category they operate in.
The first four entries describe how the work is billed, the fifth the technology category. None of them says anything about whether the deployment pays off.
Why not just use an ordinary warehouse robot?
GXO's warehouse already has automation in it. Conveyors work, racking systems work, and mobile robots have been driving around the floors for years. The question a reader asks immediately is: why add a two-legged robot.
Added, not rebuilt
The answer from this deployment is narrow, which is what makes it credible. Digit does not replace a conveyor — it works next to one, moving totes where a person used to. It is added to existing infrastructure rather than requiring that infrastructure be rebuilt.
The arithmetic
The flip side is economics. A humanoid is more expensive than a dedicated manipulator and slower than a conveyor. Its advantage is supposed to be that it needs no floor redesign — but that advantage only counts if the redesign would cost more than the robot.
Availability of people
There is a third argument, rarely said out loud: the availability of people. Moving totes for a whole shift is work that is hard to keep staffed, and turnover in such roles is high. The robot does not have to be faster than a person here — it is enough that it is available when nobody could be hired.
That is why the missing ROI is not a marginal footnote here. It is the main unresolved parameter of the whole category.
Have other humanoids passed through GXO?
Yes, and that is the strongest context for this news. GXO has tested, among others, Apollo from Apptronik and Reflex from Reflex Robotics.
| Robot | Supplier | Status at GXO |
|---|---|---|
| Digit | Agility Robotics | Commercial agreement, expanding to two sites |
| Apollo | Apptronik | Testing |
| Reflex | Reflex Robotics | Testing |
On the August earnings call GXO CEO Patrick Kelleher said plainly where the whole category stands.
We've done 45 pilots on humanoids so far. We have an additional pilot launching in Europe very shortly. We have not achieved ROI on humanoids yet. I think we are a couple of years away from that.
Patrick Kelleher, CEO of GXO Logistics, on the second-quarter 2026 earnings call.
The proportions in that remark are worth noting. Forty-five pilots says more about the category than any single press release: a logistics operator has run dozens of attempts through its floors and come out of none of them with a positive balance. Expanding the Digit deployment is therefore a choice of one supplier from that pool, not confirmation that the category has settled.
Agility's CEO put it from the other side.
Customers are moving beyond asking whether humanoid robots can work to deciding where they can scale them.
Peggy Johnson, CEO of Agility Robotics, in the press release.
Is it live yet?
Partly. At Flowery Branch it is a production deployment, at the two new sites an announcement for early 2027. The RaaS agreement signed in June 2024 covered only the Georgia site.
GXO did not say how much it is investing in the expansion or when it expects a return. Digit 5, unveiled on 15 September 2026, is not yet commercially available.
It is also worth being precise about what Digit does today: one repetitive task — moving totes and placing them on conveyors. The announcement promises a wider scope but names no specific new applications.
Why does it matter?
This is the first case of a logistics operator expanding a humanoid deployment without having worked out the return. The decision therefore rests on a trend in capability, not on a spreadsheet. For the category as a whole it signals that entry into the warehouse is happening before profitability is proven, not after.
What comes next?
- Deployments at Statesville and Lebanon are due to start in early 2027 on Digit 4, migrating to Digit 5 after its commercial launch.
- Agility has announced a Digit deployment at an AGCO agricultural equipment factory in 2027, plus a new 50,000-square-foot facility in Pittsburgh.
- Agility's combination with Churchill Capital Corp XI is expected to close in the fourth quarter of 2026, after which the company is expected to trade on Nasdaq under the ticker AGLT — the first public filings will show what this scale costs.
