Humanoid robot in the warehouse: where do things really stand in 2026?

The humanoid robot already exists in production in a warehouse, but with only one case really documented at scale: Agility Robotics' Digit at logistics provider GXO, which passed 100,000 totes moved in late 2025. Other projects — Tesla's Optimus, Unitree, Figure AI — are moving forward, but remain at the testing, demo or delayed-production stage. Meanwhile, AMRs and autonomous pallet trucks already equip warehouses under real operating conditions for pallet handling.

Digit at GXO: the most documented case of a humanoid robot in a warehouse

The only humanoid-robot deployment in a warehouse that today has public, verifiable production figures is that of Digit, Agility Robotics' bipedal robot, at global logistics provider GXO.

In June 2025, GXO signed with Agility Robotics the first multi-year RaaS-type agreement (Robot-as-a-Service, a subscription-based robot rental) specifically covering a humanoid robot in logistics (source: Agility Robotics). A few months later, on November 24, 2025, Digit passed the milestone of 100,000 totes moved under real production conditions at that GXO site (source: Robotics and Automation News).

That is a notable figure, and it is precisely what sets this case apart from other announcements in the sector: it is not a lab demo or a promotional video, but a repeated, measured activity published from a working warehouse. Amazon is also testing Digit in some of its facilities, alongside its own robotic systems such as Sequoia, Sparrow and Vulcan (source: GeekWire) — a signal that even the largest global logistics operators still treat the humanoid as a component under evaluation, not as a wholesale replacement for their existing robots.

Optimus, Unitree, Figure AI: where the other humanoid projects stand

Outside the Digit/GXO case, most announcements around the humanoid robot still amount more to industrial promise than to verified deployment.

Tesla Optimus is the most talked-about example: several trade outlets report a production delay acknowledged by the company, without detailed official confirmation and with sources that need careful cross-checking. What can be said without ambiguity is that, to date, Optimus has not published any deployment figures for a third-party warehouse comparable to Digit's 100,000 totes.

Unitree, for its part, illustrates another facet of the sector: large-scale industrial production rather than documented logistics deployment. The Chinese company delivered around 5,500 humanoid robots in 2025, close to a third of the global market, and filed for an IPO in Shanghai in March 2026, targeting between $610 million and $900 million (sources: Rest of World, CNN, CNBC). The most striking fact remains the drop in the average price of a Unitree humanoid, from $85,000 in 2023 to $25,000 in 2025 — a spectacular decrease, but one that says nothing about their actual use in warehouse logistics, where Unitree does not report a documented case comparable to GXO.

Figure AI illustrates, conversely, the financial dimension of the hype: the company raised around $1 billion in a Series C round in 2025, taking its valuation to $39 billion, while remaining essentially pre-revenue (source: Tea Hose). A valuation of this scale ahead of any significant revenue is a classic signal of a market that is still speculative.

The manufacturer, already present in the autonomous pallet truck market with the robot, is also exploring the segment with a semi-humanoid D7 and a bipedal D9 model — a technological diversification that confirms the sector's interest in the humanoid format, without a documented logistics deployment equivalent to Digit/GXO having been published to date for these models.

Why AMRs and autonomous pallet trucks remain more mature today

The humanoid robot aims for versatility: a single machine capable, in theory, of walking, grasping varied objects and adapting to an environment designed for humans. That is a legitimate ambition, but it comes at the cost of complexity — balance, grasping, safety — that the cases documented so far have not yet resolved at the scale of thousands of units in production.

For a precise, recurring need — moving pallets between a dock and a storage area — AMRs (autonomous mobile robots) and autonomous pallet trucks today address a different operational reality: fleets already installed, an established regulatory safety framework (the EN ISO 3691-4 standard) and a market growing quickly but measurably. According to Interact Analysis, the global mobile robot market is set to grow from $5 billion in 2024 to $14 billion in 2030 (+19% a year), and autonomous forklifts and pallet trucks, while accounting for only 14% of units sold, will represent 33% of the sector's revenue in 2030 — the most commercially mature segment of all mobile robotics (source: Interact Analysis).

This is not a value judgment on the humanoid robot, whose progress is real and rapid. It is a statement about timing: for the specific task of transporting pallets, a specialised machine already proven in the field remains, in 2026, a choice that is more immediately deployable than a generalist machine still being validated at scale.

Humanoid robot in the warehouse: comparison table of documented projects

Humanoid robot in the warehouse: what realistic timeline?

At the scale of a pilot site at a very large logistics provider like GXO, the humanoid robot has already moved past the demo stage. At the scale of a generalised, multi-site, multi-task deployment, current public data does not yet support that claim: the majority of announcements in the sector concern funding rounds, product launches or production targets, not verified usage volumes in warehouses.

For a French SME or mid-sized company looking to automate pallet handling today, the question, then, is not whether to wait for the generalist humanoid, but how to handle the most repetitive flow — transporting loads between dock and stock — with an already proven technology: the AMR or the autonomous pallet truck. The humanoid robot remains one to watch closely, in particular through the next figures published by GXO or other logistics operators, but it does not, in 2026, constitute an available alternative for equipping a mid-sized warehouse.

FAQ

Is the humanoid robot already used in warehouses in 2026?

Yes, but still on a limited scale, at a handful of pilot sites. The most documented case is Agility Robotics' Digit, which passed 100,000 totes moved in production at logistics provider GXO in late 2025, under the first multi-year RaaS-type agreement for a humanoid robot in logistics.

What is the difference between a humanoid robot and an AMR?

A humanoid robot is bipedal and aims for versatility, mimicking human walking and grasping in an environment not adapted for machines. An AMR (autonomous mobile robot) is a wheeled platform specialised in moving loads, which navigates using SLAM mapping. The AMR is currently more mature for pallet handling.

Is Tesla Optimus deployed in warehouses?

To date, no deployment figures for Optimus in a third-party warehouse have been published by Tesla, unlike Digit at GXO. Several trade outlets report a production delay, but this information still needs to be cross-checked against more solid sources before being treated as confirmed.

How much does a humanoid robot cost?

The only comparable public figure is Unitree's, whose average humanoid robot price fell from $85,000 in 2023 to $25,000 in 2025. This price does not necessarily cover a full logistics use case; integration costs in a warehouse are not, for their part, publicly documented.

Can a humanoid robot replace an autonomous pallet truck?

Not yet, in any documented way. For the repetitive transport of pallets, autonomous pallet trucks and AMRs today offer deployed fleets, an established regulatory framework (EN ISO 3691-4) and a measurable return on investment, while the humanoid remains concentrated on a small number of pilot sites.

Related reading

• Logistics robots: the 7 types of warehouse robot, to place the humanoid robot among the robot families already deployed in warehouses.
• AGV vs AMR: what are the differences, and which should you choose for your pallets? — understand the technology that already equips most warehouses, in contrast to the still-experimental humanoid.
• autonomous pallet truck: a complete overview of the autonomous pallet truck, an example of an already mature, deployable mobile robot requiring no infrastructure work.
• ROI of a pallet-truck robot, how to calculate a return on investment with a technology available today.
• Autonomous pallet truck and regulations: CACES, ISO 3691-4, liability — the regulatory framework already in place for AMRs, still non-existent for warehouse humanoids.

Want to automate pallet handling with an already proven technology, rather than wait for the next humanoid announcement? Easy to Carry distributes in France the autonomous pallet truck, a 2,000 kg autonomous pallet truck deployable with no infrastructure work. Discover the robot or request the free site pre-assessment: we study the feasibility for your pallet flows, with a written verdict within 48 business hours.