Agility Robotics Digit
Digit is a warehouse humanoid robot from Agility Robotics, built to move totes and feed conveyors in real logistics operations. It's offered mainly as robot-as-a-service at about $30 per hour, with a reported outright purchase price near $250,000. Digit made history as the first humanoid in a paid commercial deployment (at GXO), and by 2026 is working across multiple customer facilities.
What is the Agility Robotics Digit?
If you want to see a humanoid robot that already clocks in and does real work, Agility Robotics' Digit is the clearest example. While flashier competitors trade demo videos, Digit has quietly been moving totes on a warehouse floor - and getting paid for it.
Digit is a bipedal robot roughly the height of a person, purpose-built for logistics: taking totes from cobots, carrying them, and placing them onto conveyors. Its slightly bird-like legs are an intentional engineering choice for efficient, stable walking rather than a bid to look human.
Below we cover Digit's specs, the RaaS pricing model, its landmark GXO and Amazon work, and where it fits versus rivals like Apptronik's Apollo in 2026 - including Agility's move toward a public listing.
What is the Agility Robotics Digit robot?
Digit is a commercial humanoid robot designed by Agility Robotics for warehouse and factory logistics. It stands about 5 ft 9 in tall, weighs around 140 lb, and can lift up to roughly 35 lb (current gen), moving totes between cobots, shelves and conveyors.
Unlike humanoids that chase a fully human silhouette, Digit uses distinctive backward-bending legs tuned for efficient bipedal walking and getting into tight logistics spaces. It's orchestrated through Agility Arc, a cloud platform that coordinates fleets of Digits with existing warehouse systems.
A next-generation update, expected around late 2025 into 2026, is set to raise payload toward 50 lb, extend battery life (up to roughly 8 hours of operation) and improve manipulation.
How much does the Agility Robotics Digit cost?
Digit is sold mostly through a robot-as-a-service (RaaS) model at about $30 per hour - deliberately benchmarked against the fully loaded cost of a human worker. That means customers pay for uptime and output rather than sinking capital into hardware they then have to maintain.
For buyers who want to own units outright, a reported purchase price sits around $250,000, though most deployments to date have used RaaS. As with the whole sector, exact contract terms are negotiated per customer rather than published as a fixed list price.
The Amazon and GXO deployments - is Digit really working?
Yes, and this is Digit's strongest claim to fame. In June 2024, Digit began working at a GXO facility near Atlanta - the first time a humanoid robot was placed in a genuine commercial operation, and the first humanoid RaaS deployment. GXO later signed an industry-first multi-year agreement with Agility.
At GXO's Flowery Branch facility, Digit has moved over 100,000 totes, handling repetitive moves from cobots onto conveyors. Amazon has also tested Digit in its research and fulfillment operations. By 2026 Agility reports Digit working across around nine customer facilities - modest numbers, but they're real paid jobs, which is more than most humanoid programs can claim.
How does Digit work and how autonomous is it?
Digit operates with supervised autonomy inside structured warehouse workflows. Tasks like tote-to-conveyor moves are defined and repeated, with the Agility Arc platform handling fleet coordination, task assignment and integration with warehouse management systems.
It's not an open-ended general worker - it excels at bounded, repetitive material-handling in environments set up for it. That focus is actually why Digit reached paid deployment before more ambitious general-purpose robots: solving a narrow job well beats doing everything poorly.
Agility Digit vs Apptronik Apollo
Digit and Apollo are the two most warehouse-proven humanoids around. Digit's edge is track record - it was first into paid commercial work and has public tote numbers to show. Apollo counters with a more human-like general design, force control, and deep pockets (a $5.5B valuation in 2026 with Google and Mercedes backing).
Commercially, Agility took a notable step in June 2026 by signing a SPAC deal (with Churchill Capital XI) to list on Nasdaq, positioning itself as a pure-play humanoid public company. For a logistics buyer, both are worth evaluating; Digit is arguably the safer 'it already works here' pick today.
Is Digit worth it, and who is it for?
Digit is aimed at warehouse and 3PL operators with repetitive tote and conveyor tasks. If that's you, the RaaS model lowers the risk of adopting an emerging technology - you're paying for work, not betting a big capital outlay on v1 hardware.
It is not a consumer product and won't do arbitrary household tasks. But among humanoids you can actually put to work in 2026, Digit is one of the very few with a real operational history behind it, which makes it a sensible pilot candidate for logistics teams.
Who the Agility Robotics Digit is — and isn't — for
+Where it shines
- US-built, with local support and supply chain.
- Stated real-world payload (16 kg).
- Specs are well-sourced — high data confidence.
!Watch-outs
- Sold to businesses only — not to private buyers.
How the Agility Robotics Digit stacks up
Side by side with similar robots — the one you're reading about is highlighted.
| This robot Agility Robotics Agility Robotics Digit |  Weave Robotics Weave Isaac |  1X Technologies 1X NEO |  Foundation (Foundation Robotics) Foundation Phantom | |
|---|---|---|---|---|
| Price (US) | ~$250,000 | $7,999 – $9,999 | ~$20,000 | $100,000 – $150,000 |
| Availability | Lease / RaaS | On the market | Reservations open | Lease / RaaS |
| Height | 1.75 m | Unknown | ~1.68 m | 1.75 m |
| Payload | 16 kg | Unknown | Carry 25 kg | 40 kg dual-arm |
| DoF | 16 | Isaac | Hands | 29 |
| Buyer | Enterprise | Individuals | Individuals | Enterprise |
| Made in | US | US | US | US |
What the Agility Robotics Digit costs — and how to get one
Mainly RaaS (~$30/hr); reported outright pilot purchase ~$250k.
Open - Deploy Digit / sales contact
Leasing / RaaS: Yes - primarily RaaS; GXO industry-first multi-year RaaS (Jun 2024)
Every spec, grouped & sourced
Tap a group to expand. Values are labelled official vs third-party; unknowns are omitted, not guessed.
Body & movement▾
Power & endurance▾
Hands & senses▾
Brains & control▾
Claims vs evidence
Autonomous task execution via Agility Arc cloud; can self-stand after fall; autonomy level unverified
Where it's working
GXO (SPANX, Flowery Branch GA - 100k+ totes), Toyota MMC Canada, Schaeffler, Mercado Libre
Intended use cases: Warehouse/factory logistics: tote handling, AMR-to-conveyor moves, bulk material handling, sortation
Should you consider the Agility Robotics Digit?
Digit is the humanoid that actually shows up for work - first into paid deployment, real tote counts at GXO, and a RaaS model that de-risks adoption. It's a narrow, warehouse-focused robot rather than a do-anything android, but that focus is exactly why it's ahead. For logistics operators, it's one of the safest humanoid pilots to run in 2026.
Score reflects stated criteria (availability, data confidence, US relevance) — not hands-on testing or user ratings. We're an independent review & comparison site; we don't sell robots. *Financing figure is illustrative.
How we know this
Data confidence 78/100 · last verified 2026-06-23. Specs are labelled official vs third-party and dated; unknowns are omitted, not guessed.
Agility Robotics Digit — your questions, answered in full
How much does the Agility Robotics Digit cost?+
Mainly offered as robot-as-a-service at about $30 per hour, benchmarked to a human worker's loaded cost. A reported outright purchase price is around $250,000, though most deployments use the RaaS model.
What can the Digit robot do?+
It handles warehouse logistics - moving totes from cobots and shelves onto conveyors, plus bulk material handling and sortation. Current payload is up to about 35 lb, rising toward 50 lb in the next generation.
Is Digit actually deployed in real warehouses?+
Yes. Digit began paid work at a GXO facility near Atlanta in June 2024 - the first commercial humanoid deployment - and has moved 100,000+ totes. By 2026 it's working across roughly nine customer facilities.
Did Amazon use the Digit robot?+
Amazon has tested Digit within its research and fulfillment operations. The best-documented commercial, contracted deployment, however, is with GXO Logistics.
What are Digit's specifications?+
About 5 ft 9 in tall, ~140 lb, current payload up to ~35 lb, and battery life up to roughly 8 hours in newer configurations. It's coordinated by Agility's cloud platform, Agility Arc.
Who makes the Digit humanoid robot?+
Agility Robotics. In June 2026 the company signed a SPAC deal with Churchill Capital XI to list on Nasdaq as a pure-play humanoid robotics company.
Digit vs Apptronik Apollo - which is better for warehouses?+
Digit has the longer paid track record and public deployment numbers. Apollo brings a more general design, force control and heavy funding. Both are legitimate; Digit is the more proven 'already working' choice today.
Can I buy a Digit robot?+
Enterprises can, via Agility Robotics - either through RaaS (about $30/hr) or a reported ~$250k purchase. It's not sold to consumers, and terms are negotiated per customer.
Peer-reviewed research on the Agility Robotics Digit
Published academic work — peer-reviewed papers and preprints — involving this robot or its maker's systems, indexed in scholarly databases (OpenAlex, arXiv). Not blog posts or press write-ups.
- Reinforcement Learning for Robust Parameterized Locomotion Control of Bipedal RobotsZhongyu Li, Xuxin Cheng, Xue Bin Peng et al. · 2021 · IEEE ICRA 2021 · DOI 10.1109/ICRA48506.2021.9560769
- Sim-to-Real Learning of All Common Bipedal Gaits via Periodic Reward CompositionJonah Siekmann, Yesh Godse, Alan Fern et al. · 2021 · IEEE ICRA 2021 · DOI 10.1109/ICRA48506.2021.9561814
- Feedback Control For Cassie With Deep Reinforcement LearningZhaoming Xie, Glen Berseth, Patrick Clary et al. · 2018 · IEEE/RSJ IROS 2018 · DOI 10.1109/IROS.2018.8593722
- Real-World Humanoid Locomotion with Reinforcement LearningIlija Radosavovic, Tete Xiao, Bike Zhang et al. · 2024 · Science Robotics · DOI 10.1126/scirobotics.adi9579
Academic references are factual citations from open scholarly indexes; listing is not endorsement.