GXO Adds Agility Robotics Digit Humanoids at Two More US Warehouses

GXO Logistics said on October 26, 2026 that it will deploy Agility Robotics' Digit humanoid at warehouses in Statesville, North Carolina, and Lebanon, Tennessee, with work expected to begin in early 2027. The move extends a partnership that started as a proof of concept in late 2023 and became a multi-year commercial agreement in June 2024 at a site in Flowery Branch, Georgia.

GXO chief executive Patrick Kelleher told analysts on the company's second-quarter earnings call in August 2026 that the company had run 45 humanoid pilots and had not achieved a return on investment on any of them, a position that sits uneasily beside a fresh expansion. GXO describes itself as the world's largest pure-play contract logistics provider, with more than 150,000 team members across more than 1,000 facilities totaling more than 200 million square feet, so even a small deployment carries symbolic weight.

Humanoid robots have become one of the more heavily funded corners of industrial automation. Agility Robotics, based in Salem, Oregon, is preparing to list on Nasdaq through a business combination with Churchill Capital Corp XI, a special purpose acquisition company, under the ticker AGLT. Its investors include NVIDIA, Amazon, SoftBank Vision Fund 2, Schaeffler, Foxconn, Abico, DCVC and Playground Global, a roster that explains why each announcement is read as a signal about the category.

Warehouse operators have spent a decade layering automation into facilities: conveyors, sortation systems, autonomous mobile robots and, more recently, fixed robotic arms. Humanoids are the newest and least proven layer, marketed as flexible enough to handle tasks that fixed machinery cannot reach without rebuilding a building. The economics are usually framed as robots-as-a-service, a recurring fee rather than an outright hardware purchase, and that structure is central to how GXO and Agility describe their work together.

Key Facts

The new agreement covers e-commerce fulfillment at the Statesville and Lebanon sites, where Digit will take on what the companies call highly repetitive material handling tasks. DC Velocity reported on October 26, 2026, that the deployments follow the pattern set in Georgia, where Digit works alongside existing automation to move totes and place them onto conveyors, handling inventory for the women's clothing brand Spanx. GXO and Agility call Flowery Branch the industry's first formal commercial deployment of humanoid robots in live operations and the first robots-as-a-service agreement for humanoids.

Since that deployment began, Digit has moved more than 100,000 totes at Flowery Branch. Agility Robotics said in its announcement on October 26, 2026, that work at the two new sites will start on Digit 4 and migrate to Digit 5 once that model is commercially available. Digit 5 was unveiled on September 15, 2026, is designed for cooperatively safe work close to people, and is expected to reach early customer access in the first half of 2027 and general availability by the end of that year.

Across all commercial operations and customer programs, Digit has accumulated more than 65,000 hours of real world operation, according to the companies. Agility runs RoboFab, a 70,000 square foot plant in Salem, Oregon, designed for annual capacity of up to 10,000 robots, plus its Agility Arc cloud platform for managing fleets and a 50,000 square foot site in Pittsburgh. Neither company disclosed the number of robots, the contract value or the service fee for the new locations, and GXO did not say how much it is investing or when it expects a return.

The August earnings call remains the sharpest public statement of the economics. The Motley Fool published a transcript on August 11, 2026, in which Kelleher said that in production, zero humanoids would be deployed at GXO this year, that the company had run 45 humanoid pilots so far, and that it had not yet achieved ROI on humanoids, which he placed a couple of years away. He noted a further pilot launching in Europe shortly. On the same call GXO reported second-quarter revenue of $3.4 billion, up 4 percent year over year and 3.4 percent organically, adjusted EBITDA of $219 million at a 6.4 percent margin, adjusted earnings per share of $0.59 and new business wins of $410 million, alongside a sales pipeline of $2.7 billion in July and a plan to deploy 20,000 robots across its network during the fiscal year.

Agility used the same day to widen the picture. It disclosed three additional enterprise manufacturing customers in paid programs and a pipeline of more than 30 prospective customers across automotive manufacturing, logistics, e-commerce, agricultural equipment, consumer products and defense. Named customers include AGCO, which announced a commercial agreement on October 20, 2026 for a Digit pilot at its Hesston, Kansas plant expected to begin in 2027, plus Toyota Motor Manufacturing Canada, Schaeffler, Mercado Libre and GXO. Agility also cites more than $300 million in committed multi-year orders for Digit 5, a figure it describes as potential multi-year value subject to contractual milestones rather than current period revenue.

Analysis

The obvious reading is contradiction: a company that says humanoids have not paid back is buying more of them. MillBrief argued in a brief published on October 27, 2026, that the two facts are compatible only if the expansion is treated as continued learning rather than a proven return. GXO has also tested humanoids from Apptronik and Reflex Robotics, which points to a portfolio approach to evaluating the category rather than a single vendor bet. What this really means is that the deployment is an option being purchased, not a verdict being delivered, and the missing numbers are exactly the ones that would settle the argument.

Kelleher's own framing supports that reading. He said humanoids will absolutely feature in GXO's solutions while placing production scale about two years out, which is a statement about timing, not rejection. The Georgia tote count is a real operational datapoint, more than 100,000 totes moved, but it says nothing about cost per tote, uptime, intervention rates or labor hours avoided. A tote count without those figures is a throughput anecdote, and the silence on robot counts and fees leaves the same gap.

The industry pattern is similar. MillBrief noted that Schaeffler has built a humanoid lighthouse factory in Taicang, China, covering about 7,500 square meters, with six humanoids at five stations assembling linear and rotary actuators for other humanoids, and that Schaeffler reported roughly 400 million euros in global incoming humanoid orders after two new Greater China customers in the third quarter while publishing no output, cycle time or uptime data for the line. The bigger picture here is that the sector is still reporting orders, pilots and hours rather than unit economics.

For a mid-size plant, the practical conclusion is patience with structure. MillBrief's advice is to ask for rate per hour at a running site, availability share and annual service cost per robot, then compare those against conveyors or mobile robots on total cost of ownership. That is a procurement checklist, not a rejection. The expansion is best read as evidence that a very large operator with 45 pilots behind it will keep paying for optionality, which is different from evidence that payback has arrived.

Why It Matters

The warehouse labor market and the automation supply chain are both watching this pair of announcements because GXO is a bellwether buyer. If the largest pure-play contract logistics provider keeps expanding humanoid deployments while publicly doubting the return, other operators get cover to run small pilots without committing to fleet scale. That dynamic favors vendors with live reference sites and penalizes those selling projections.

It also matters for the public market story around Agility. The company is heading toward a Nasdaq listing under AGLT through Churchill Capital Corp XI, with the business combination expected to close in the fourth quarter of 2026, and investors will read commercial traction through exactly the lens Kelleher described. Customer logos, hour counts and order commitments are the visible metrics, but a signed expansion with undisclosed robot counts and fees does not resolve the ROI question the chief executive himself put on the record in August.

Finally, the disclosure gap is itself the story: companies in this category routinely announce deployments without the figures that would let an outsider test the economics, and Kelleher's unusually candid remarks give the market a benchmark the announcements neither contradict nor confirm.

Next Up

The near-term checkpoints are concrete. AGCO's Digit pilot at Hesston, Kansas is expected to begin in 2027, the Statesville and Lebanon deployments are expected to start in early 2027 on Digit 4, and Digit 5 is slated for early customer access in the first half of 2027 with general availability by the end of that year. Agility's combination with Churchill Capital Corp XI is expected to close in the fourth quarter of 2026.

The number to watch is not the tote count but the first credible cost per hour at a running site. Until an operator publishes that, or until Kelleher revises his two-year timeline, the expansion stands as a bet placed by a buyer who says the payoff is still ahead.