Agility Digit
Last reviewed
Sources
44 citations
Review status
Source-backed
Revision
v6 ยท 4,629 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
44 citations
Review status
Source-backed
Revision
v6 ยท 4,629 words
Add missing citations, update stale details, or suggest a clearer explanation.
| Agility Digit | |
|---|---|
 | |
| General information | |
| Manufacturer | Agility Robotics |
| Country of origin | United States |
| Year introduced | 2017 (v1); 2023 (commercial version) |
| Status | In production |
| Price | ~$250,000 USD (also available via RaaS) |
| Availability | Available (2026 production fully committed) |
| Website | agilityrobotics.com |
Agility Digit is a bipedal humanoid robot developed by Agility Robotics (rebranded as Agility in March 2026), an American robotics company headquartered in Salem, Oregon. Designed primarily for warehouse logistics and material handling, Digit is widely regarded as the most commercially deployed humanoid robot in the world as of 2026: by November 20, 2025 its fleet had moved more than 100,000 totes in a live, revenue-generating fulfillment operation at GXO Logistics in Flowery Branch, Georgia.[20][40] The robot features a distinctive backward-bending leg design inspired by ostrich biomechanics, purpose-built gripper hands that achieve a greater than 99% grasp success rate on warehouse totes, and a sensor suite that includes LiDAR, Intel RealSense depth cameras, and an IMU.[13] Standing 175 cm (5 ft 9 in) tall and weighing approximately 65 kg (140 lb), Digit can carry payloads up to 16 kg (35 lb) and walk at speeds up to 1.5 m/s.
Digit is the first humanoid robot to work full-time in commercial warehouse and manufacturing settings. It has been commercially deployed at facilities operated by GXO, Amazon, Schaeffler, Mercado Libre, and Toyota Motor Manufacturing Canada. "There will be many firsts in the humanoid robot market in the years to come, but I'm extremely proud of the fact that Agility is the first with actual humanoid robots deployed at a customer site, generating revenue, and solving real-world business problems," said Agility Robotics CEO Peggy Johnson when the first GXO commercial agreement was announced in June 2024.[19]
Digit is a human-sized, bipedal humanoid robot built specifically for repetitive material-handling work in warehouses and factories rather than as a general-purpose machine. It is purpose-engineered to move totes, boxes, and bins through workflows designed for human bodies, picking containers on and off autonomous mobile robots, loading them onto conveyors, and stacking them at floor locations.[20] Unlike many competitors pursuing broad, dexterous, multi-purpose humanoids, Agility deliberately narrowed Digit's design, grippers, and software to the logistics tasks that dominate fulfillment centers, which is the strategy that let it reach paying commercial deployments ahead of most rivals.[37]
Agility Robotics was founded in November 2015 as a spin-off from Oregon State University's Dynamic Robotics Laboratory. The company was co-founded by Jonathan Hurst, Damion Shelton, and Mikhail Jones.[1] Hurst, a professor of robotics at Oregon State, had spent over a decade researching dynamic stability and locomotion in bipedal robots. Shelton, who met Hurst during their Ph.D. studies at Carnegie Mellon University in the early 2000s, brought entrepreneurial experience from co-founding and exiting 3D imaging startup threeRivers 3D in 2012. Jones, a graduate student of Hurst's at OSU, served as the lead developer of the control software behind the lab's earlier robot platforms.[2]
The company's technical foundation rests on two predecessor robots built at Oregon State: ATRIAS and Cassie.
ATRIAS was developed in Oregon State's Dynamic Robotics Laboratory as a research platform for studying bipedal locomotion. It was the first robot to reproduce human-like walking gait dynamics and implement spring-mass walking, a model that treats the leg as a simple spring and mass system to achieve energy-efficient locomotion.[3] While ATRIAS demonstrated groundbreaking science, it was not a practical machine for any application beyond laboratory demonstration. The principles validated by ATRIAS, particularly around dynamic balance and compliant leg design, became the scientific bedrock upon which Agility Robotics built its commercial products.
Cassie, introduced in 2016, represented the next evolution of bipedal robotics from the Oregon State lab. Developed with a 16-month, $1 million grant from the Defense Advanced Research Projects Agency (DARPA), Cassie was a legless biped with no upper body, torso, or perception systems.[4] Despite this minimal form factor, Cassie proved remarkably capable. The robot was sold as a research platform to academic laboratories and corporate innovation departments starting in 2016.
Cassie went on to achieve two notable milestones. In July 2021, using machine learning to control a running gait on outdoor terrain, Cassie completed a 5K run on Oregon State's campus untethered and on a single battery charge, finishing in just over 53 minutes.[5] In September 2022, Cassie set a Guinness World Record for the fastest 100-meter dash by a bipedal robot, clocking 24.73 seconds at OSU's Whyte Track and Field Center. The robot started from a standing position and returned to standing after the sprint, with no falls and no external sensors (essentially running blind).[6]
Digit v1 (2017): The first version of Digit added a torso, arms, and a full perception system to the Cassie leg platform, creating a complete humanoid form factor. Digit v1 was sold to academic labs and research institutes as a research platform.[7]
Digit v2 (2019): The second version introduced nimbler feet with a roll degree of freedom, allowing the robot to resist lateral forces without needing to sidestep. The update also included better perception and fully custom electronics.[8]
Digit v3 (2020): Changes from v2 to v3 were primarily incremental, focusing on bug fixes and regulatory upgrades such as full battery certification. In January 2020, Ford Motor Company became the first commercial customer, purchasing two Digit v3 units to explore last-mile delivery in combination with autonomous vehicles.[9]
Digit v4 (2023): Unveiled at the ProMat 2023 trade show in Chicago, the fourth generation represented a major redesign for industrial deployment. Key changes included the addition of a head-like structure to improve human-robot interaction, new gripper hands designed specifically for tote manipulation, a streamlined body optimized for warehouse environments, and enhanced computing and perception systems. This was the version that transitioned Digit from a research platform to a commercial product.[10]
Ongoing updates (2025-2026): Agility has continued to improve Digit through iterative hardware and software updates, including extended battery life (up to four hours), autonomous docking for charging, expanded grasping angles with new robust limbs and end effectors, Category 1 (CAT1) safety stop functionality, and integration with autonomous mobile robot (AMR) platforms from MiR and Zebra Technologies.[11]
Digit's most distinctive physical feature is its backward-bending "bird legs," inspired by ostrich biomechanics. This leg geometry, carried forward from the Cassie platform, provides several advantages in warehouse environments. The backward-bending knees give Digit excellent dynamic balance, the ability to crouch low to reach into tight spaces, and a compact folding profile that allows the robot to fit into confined areas that would be difficult for a conventionally proportioned humanoid to navigate.[12]
The robot lacks a traditional head in the conventional sense. Early versions had no head at all; Digit v4 introduced a head-like structure primarily as a human-robot interaction focal point, giving nearby workers a visual cue to understand what Digit is looking at and where it intends to move. The body is covered in teal-colored panels that serve both protective and aesthetic purposes.
Digit's hands are purpose-built gripper-style end effectors optimized for grasping standard warehouse containers such as totes, boxes, and bins. Rather than attempting to replicate human hand dexterity with multi-fingered designs, Agility chose specialized grippers that achieve greater than 99% grasp success rates on warehouse containers.[13] The grippers can handle payloads up to 16 kg (35 lb), sufficient for the vast majority of warehouse material handling tasks. The end effectors are interchangeable, allowing adaptation to different use cases.
| Category | Specification | Details |
|---|---|---|
| Physical | Height | 175 cm (5 ft 9 in) |
| Physical | Weight | ~65 kg (140 lb) |
| Physical | Form factor | Bipedal humanoid, backward-bending legs |
| Degrees of freedom | Total | 16 (core); up to 30 (with all joints) |
| Degrees of freedom | Per leg | 5-6 |
| Degrees of freedom | Per arm | 3-4 |
| Mobility | Max walking speed | 1.5 m/s (5.4 km/h; 3.4 mph) |
| Mobility | Stair climbing | Yes |
| Mobility | Incline navigation | Yes |
| Mobility | Self-recovery from falls | Yes |
| Manipulation | Payload capacity | 16 kg (35 lb) |
| Manipulation | Hand type | Specialized grippers (interchangeable) |
| Manipulation | Grasp success rate | >99% on standard totes |
| Power | Battery type | Custom 1.2 kWh lithium-polymer |
| Power | Runtime | Up to 4 hours (with 2025 updates) |
| Power | Charging | Autonomous docking |
| Compute | Processors | Dual Intel i7 multi-thread CPUs |
| Compute | Expansion | Payload bay for NVIDIA Jetson or Intel NUC |
| Sensors | Vision | 4x Intel RealSense depth cameras, RGB cameras |
| Sensors | LiDAR | Yes |
| Sensors | IMU | MEMS IMU |
| Sensors | Proprioception | Absolute and incremental encoders |
| Sensors | Force sensing | Force sensors in each arm |
| Actuators | Type | Brushless DC motors with custom transmissions |
| Actuators | Peak joint torque | Up to 120 Nm |
| Connectivity | Interfaces | WiFi, Ethernet, 4G LTE |
| Connectivity | Software compatibility | ROS compatible, Linux-based real-time system |
| Safety | Features | CAT1 stop, Safety PLC (PLd), on-robot E-stop, FSoE |
Digit's locomotion system supports walking forward, backward, and sideways, turning in place, crouching, navigating inclines and stairs, and stepping over obstacles. The robot can balance dynamically using both its arms and feet when bumped or disturbed, and it can autonomously recover from falls.[14] This robustness stems from the dynamic walking principles originally validated in the ATRIAS and Cassie platforms.
The sensor suite, combining LiDAR, four Intel RealSense depth cameras, and an IMU, enables Digit to perceive and navigate dynamic environments alongside human workers. Force sensors in each arm enable compliant manipulation, allowing the robot to handle totes of varying weights without damaging contents.
Digit runs on a Linux-based real-time operating system. The robot's behavior is generated through a combination of classical control methods, teleoperated demonstrations, reinforcement learning, and extensive simulation training. Agility emphasizes a philosophy of "there is no scale without safety," using this blended approach to validate behaviors before deployment in live environments.[15]
A large share of that training happens in simulation. Agility builds Digit's whole-body controller on NVIDIA's Isaac Sim and Isaac Lab frameworks, using reinforcement learning across millions of parallel episodes and validating the resulting policies through a MuJoCo pipeline before they reach a physical robot.[41] The approach compresses the development loop from weeks to days. "Isaac Sim running on NVIDIA GPUs lets us simulate years of real-world learning for Digit in just hours," said Pras Velagapudi, Chief Technology Officer of Agility Robotics. "That simulation speedup means we can train for all the conditions we might see on the factory floor."[41]
The robot is managed through Agility Arc, the company's cloud-based fleet management platform (discussed in detail below).
Agility Arc is a cloud automation platform for deploying and managing fleets of Digit robots. Launched in 2024, it is the first humanoid fleet management system to be successfully used in a commercial environment.[16]
Agility Arc provides operators with:
| Feature | Description |
|---|---|
| Fleet monitoring | Real-time visibility into robot status across all deployed units |
| KPI tracking | Uptime, throughput, Mean Time Between Incidents (MTBI), and productivity metrics |
| Task management | Assignment and scheduling of robot workflows |
| Facility mapping | Configuration of workcell layouts and navigation paths |
| AMR integration | Dispatching and coordination with autonomous mobile robots (MiR, Zebra) |
| Enterprise integration | APIs for WMS, WES, MES, and PLC systems |
| Remote maintenance | Over-the-air software updates and remote diagnostics |
| Charger management | Monitoring of autonomous docking and charging stations |
The platform enables Digit to autonomously dispatch AMRs to deliver items to packout stations and other locations, reducing the need for constant human intervention. This level of coordination between humanoid robots and existing warehouse automation infrastructure is a key differentiator for Agility's commercial offering.[17]
Digit's most significant commercial deployment has been with GXO Logistics, the world's largest pure-play contract logistics provider. In early 2024, GXO began a proof-of-concept pilot at a SPANX facility in Flowery Branch, Georgia, where GXO manages warehouse operations for the global womenswear brand's direct-to-consumer and in-store orders.[18]
On June 5, 2024, Digit was formally deployed in commercial operations at the GXO facility near Atlanta, marking the first time a humanoid robot entered the workforce in a commercial warehouse setting. On June 27, 2024, GXO signed the industry's first multi-year commercial agreement with Agility Robotics, structured as a Robots-as-a-Service (RaaS) contract. This represented both the first formal commercial deployment and the first RaaS agreement for humanoid robots in the logistics industry.[19] "Our R&D approach is to partner with developers all over the world to help them build and validate practical use cases that improve the working environment for our employees while optimizing operations for our customers," said Adrian Stoch, Chief Automation Officer at GXO.[19]
At the GXO facility, Digit robots handle tasks including picking totes from autonomous mobile robots, transferring goods onto conveyor systems, and stacking containers at various floor locations. On November 20, 2025, Agility announced that the Digit fleet at GXO's Flowery Branch facility had moved more than 100,000 totes in live commerce operations, demonstrating consistent performance across thousands of cycles in varying conditions.[20][40] The company framed the milestone as proof of industrial reliability rather than a demo, noting that automation "can't maintain performance across thousands of cycles and handle varying conditions over an extended deployment" is "a liability, not an asset."[40] Supply & Demand Chain Executive magazine recognized GXO as the overall winner of the 2024 Top Supply Chain Projects awards for this deployment.
Amazon began testing Digit prototypes at its robotics research and development facility near Seattle in October 2023. The initial application involved "tote consolidation," where Digit robots organized and recycled empty storage bins after inventory had been completely picked from them.[21] This task is highly repetitive and was identified by Amazon as a strong candidate for robotic automation.
Amazon participated in Agility's funding through its Industrial Innovation Fund and has broadened the partnership over time. The company has stated that Digit's size and shape are well suited for buildings designed for humans, and that it sees significant opportunity in scaling mobile manipulator solutions that can work collaboratively alongside employees.[22]
Schaeffler, the German motion technology company, has both invested in Agility Robotics and committed to deploying Digit robots at its global facilities.[25] On November 13, 2024, Agility announced that Schaeffler had become a strategic investor and customer, planning to deploy Digit across its network of roughly 100 plants worldwide by 2030.[42] "We are excited to welcome Schaeffler as a customer and investor, and look forward to delivering value in their operations with Digit," Peggy Johnson said of the agreement.[42] By 2026, Digit was performing high-tolerance manufacturing work at Schaeffler's plant in Cheraw, South Carolina, marking its move from tote handling into precision production tasks.[41]
On December 10, 2025, Latin American e-commerce giant Mercado Libre announced a commercial agreement with Agility Robotics to deploy Digit robots at its fulfillment center in San Antonio, Texas. The robots handle repetitive and physically strenuous tasks such as moving totes and transporting materials within the facility. The partnership aims to explore how automating tasks that are difficult to recruit for could increase ergonomic safety, reduce labor gaps, and improve productivity. Mercado Libre and Agility plan to expand humanoid deployment to additional warehouses across Latin America.[23]
On February 19, 2026, Toyota Motor Manufacturing Canada (TMMC) announced a commercial agreement with Agility Robotics following a successful pilot program. The pilot involved three Digit robots and included development, proof-of-technology, and onsite evaluation phases. TMMC plans to deploy seven additional robots to load and unload totes from automated tuggers on its automotive production lines, with potential for further expansion. This deployment marks the first commercial use of humanoid robots in Canadian automotive manufacturing and represents Digit's expansion beyond warehouse logistics into automotive production.[24]
Additional partnerships include collaborations with NVIDIA for simulation, training, and AI acceleration (announced March 2025); Zion Solutions (May 2024) as a systems integrator for full-cycle deployment; and Ricoh USA (September 2024) for installation and customer support across North America.[26]
Agility Robotics opened RoboFab in Salem, Oregon in late 2023, billing it as the world's first factory dedicated to manufacturing humanoid robots. The facility spans 70,000 square feet and is located approximately 30 miles from Agility's original engineering center in Corvallis, Oregon.[27]
The company has taken a measured approach to scaling production. In its first year, RoboFab targeted output of eight units per shift, allowing the team to refine manufacturing processes while avoiding waste. The facility is designed to scale to more than 10,000 robots per year at full capacity, at which point it would employ over 500 workers.[28]
Agility also operates engineering and validation facilities in Pittsburgh, Pennsylvania and AI development operations in Fremont, California. As of April 2026, the company has stated that all of its 2026 production is fully committed to existing customers, including GXO, Amazon, Schaeffler, Mercado Libre, and Toyota Motor Manufacturing Canada.[29][43]
Digit became the first FCC-approved and NRTL (Nationally Recognized Testing Laboratory) certified humanoid robot in 2025. The NRTL certification, recognized by OSHA, was obtained through a live field evaluation at a customer fulfillment site. Testing covered compliance with ANSI/RIA R15.08 (industrial mobile robot safety) and ISO 13849 (safety of machinery control systems).[30]
The robot's safety architecture includes several layers of protection:
| Safety feature | Function |
|---|---|
| Category 1 (CAT1) stop | Maintains power during deceleration for smooth, controlled halting |
| Safety PLC | Provides Performance Level d (PLd) safety functions |
| On-robot E-stop | Emergency stop button for immediate halt of all movements |
| Wireless teach pendant | Remote control with integrated E-stop |
| FSoE (FailSafe over EtherCAT) | Secure communication protocol for emergency signals |
As of early 2026, Digit operates in sectioned-off zones within warehouses, with physical separation from human workers. Agility is pursuing ISO functional safety certification, targeted for mid-to-late 2026, that would make Digit the first humanoid robot cleared to work cooperatively alongside people without physical barriers. The next-generation model is expected to include cooperative safety capabilities and an increased payload capacity of up to 23 kg (50 lb).[31]
In March 2024, Agility Robotics appointed Peggy Johnson as Chief Executive Officer. Johnson previously served as CEO of Magic Leap, where she led the company's pivot from consumer to enterprise augmented reality products, and spent six years as Executive Vice President of Business Development at Microsoft, reporting directly to CEO Satya Nadella. Before Microsoft, she held leadership positions across engineering, sales, and business development during a 24-year career at Qualcomm.[32] On joining, Johnson described Agility as a company that "stands apart for having resolutely, steadily, and remarkably made a human-centric robot that actually works, and in demanding customer environments."[32]
Co-founder Damion Shelton transitioned from CEO to President upon Johnson's appointment and later took on the role of Chief Engagement Officer, focusing on commercial and investor engagements. Co-founder Jonathan Hurst serves as Chief Robot Officer. Other key executives include Melonee Wise (Chief Product Officer), Pras Velagapudi (Chief Technology Officer), Daniel Diez (Chief Business Officer), and Jennifer Hunter (Chief Financial Officer).[33]
In March 2026, the company rebranded from "Agility Robotics" to simply "Agility," signaling its ambition to expand beyond robot hardware into the broader software, services, and infrastructure needed to scale humanoid automation across industries.[34]
Agility Robotics has raised approximately $683 million in total funding across multiple rounds.
| Round | Date | Amount | Lead investors | Notable participants |
|---|---|---|---|---|
| Seed | October 2016 | $792,000 | Robotics Hub | N/A |
| Series A | March 2018 | $8 million | Playground Global | Sony Innovation Fund |
| Series A (extensions) | 2019-2020 | ~$20 million | Multiple | DCVC, TDK Ventures, Safar Partners |
| Series B | April 2022 | $150 million | DCVC, Playground Global | Amazon Industrial Innovation Fund |
| Series C (tranches) | 2024-2025 | ~$462 million | WP Global Partners | SoftBank, Amazon, DCVC, Playground Global, Schaeffler |
The company's Series C funding, closed in 2025, valued Agility Robotics at approximately $2.1 billion (pre-money). Major investors across all rounds include Amazon (through its Industrial Innovation Fund), DCVC, Playground Global, Sony Innovation Fund, SoftBank, Schaeffler, WP Global Partners, TDK Ventures, and Safar Partners.[35][36]
Digit operates in the rapidly growing humanoid robot market, which was valued at approximately $2.9 billion in 2025 and is projected to reach $15.3 billion by 2030. Agility's primary competitors include several well-funded companies pursuing different approaches to humanoid robotics.
| Company | Robot | Approach | Key differentiator |
|---|---|---|---|
| Agility Robotics | Digit | Warehouse logistics specialist | Most commercially deployed; RaaS model; dedicated factory |
| Figure AI | Figure 02 | General-purpose humanoid | $39B valuation; BMW factory deployment; OpenAI partnership |
| Tesla | Optimus | General-purpose humanoid | Internal factory testing; projected $30,000 price point |
| Boston Dynamics | Atlas (electric) | Research and industrial | Hyundai ownership; Google DeepMind AI partnership |
| Apptronik | Apollo | General-purpose industrial | NASA heritage; Mercedes-Benz partnership |
| Sanctuary AI | Phoenix | General-purpose with dexterous hands | Carbon (AI control system); Magna partnership |
Agility differentiates itself from competitors in several ways. While many humanoid robot companies focus on building general-purpose robots with dexterous hands and broad capabilities, Agility has deliberately specialized in warehouse logistics, optimizing Digit's design, grippers, and software for the specific tasks found in fulfillment centers and manufacturing lines. This focused approach has allowed Agility to achieve commercial deployments and revenue ahead of most competitors.[37]
Digit's RaaS business model is another differentiator. Rather than requiring customers to make large upfront capital expenditures, Agility offers monthly recurring fee structures that cover hardware, dedicated workcells, and software. The company estimates customer ROI under two years when benchmarked against the fully loaded cost of human workers at $30 per hour.[38]
However, Agility faces challenges. Tesla's projected Optimus price point of approximately $30,000 per unit, if achieved at scale, could significantly undercut Digit's estimated $250,000 price tag. Figure AI's valuation of $39 billion (as of September 2025) reflects investor confidence in a more general-purpose approach that could eventually address a wider range of tasks. Boston Dynamics, backed by Hyundai's manufacturing expertise and Google DeepMind's AI capabilities, brings decades of robotics experience and substantial resources; Hyundai has committed its entire 2026 electric-Atlas production run to internal training and to Google DeepMind, and aims to build up to 30,000 Atlas units a year by 2028.[39][44]
Digit's primary applications center on warehouse logistics and manufacturing, with the following demonstrated use cases:
| Use case | Description |
|---|---|
| Tote picking and placement | Moving totes from AMRs to conveyors and between floor locations |
| Tote stacking and unstacking | Organizing containers at staging areas |
| Tote recycling | Clearing and organizing empty bins after inventory picking |
| Goods-to-person sorting | Sorting items for outbound order fulfillment |
| AMR loading and unloading | Transferring goods between AMRs and workstations |
| Palletizing | Stacking items onto pallets for shipment |
| Putwall operations | Automated sorting into put-to-light walls |
| Tugger loading/unloading | Loading and unloading totes from automated guided vehicles |
| Nesting | Organizing and nesting empty containers for storage |
Agility has positioned Digit for tasks that are repetitive, physically taxing, and difficult to recruit human workers for. The company has not pursued applications such as household assistance, healthcare, or outdoor navigation, maintaining a clear focus on industrial environments.