Agility Digit
Home > Hardware > Humanoids > Digit
Terminology
| Abbreviation | Definition |
|---|---|
| DoF | Degrees of Freedom |
| AMR | Autonomous Mobile Robot |
| RaaS | Robot-as-a-Service |
| PoC | Proof of Concept |
| VLA | Vision-Language-Action |
Overview
| Item | Content |
|---|---|
| Manufacturer | Agility Robotics |
| Headquarters | Albany, Oregon / Pittsburgh, Pennsylvania |
| Founded | 2015 (Oregon State University spinoff) |
| Market | Industrial, logistics, warehouse automation |
| Key Partners | Amazon, GXO Logistics |
Key Significance
Agility Robotics’ Digit is one of the first humanoid robots commercially deployed in logistics environments. In Fall 2025, it demonstrated the industrial practicality of humanoid robots by moving over 100,000 totes at GXO logistics centers 1. The key competitive advantage is that it can operate on stairs, narrow corridors, and irregular terrain where wheel-based AMRs cannot go, and can be deployed in existing human-designed facilities without separate infrastructure changes.
Note: Agility Robotics markets Digit as the “industry’s first commercially deployed humanoid” 2, though this is based on full-time commercial operations in logistics/warehouse sector. Competitors are also conducting various forms of pilot deployments.
Company History: From Oregon State to RoboFab
Founders and Origins
| Founder | Role | Background |
|---|---|---|
| Jonathan Hurst | Chief Robot Officer | Oregon State University Professor, CMU PhD |
| Damion Shelton | Chairman of the Board | CMU PhD, Hurst’s classmate |
| Mikhail Jones | VP of Software | Oregon State graduate |
Agility Robotics was spun off from Oregon State University’s Dynamic Robotics Lab in 2015 3. Founder Jonathan Hurst obtained his robotics PhD from CMU, then researched the physics of legged locomotion at OSU and co-founded the OSU Robotics Institute.
Cassie: Digit’s Predecessor
Cassie was Agility’s first robot, designed inspired by the leg mechanics of ostriches and cassowaries.
| Item | Content |
|---|---|
| Development Funding | DARPA $1M grant, 16 months 4 |
| Features | Pure bipedal robot without upper body/cognitive system |
| Record | World’s first 5K run completion, 2021, approximately 53 minutes 5 |
| Use | Sold to universities/research institutions as research platform |
“We weren’t trying to replicate the appearance of animals, but rather replicating the techniques animals use for agile, efficient, and robust locomotion.” - Jonathan Hurst 4
Digit Version History
| Version | Period | Key Features | Source |
|---|---|---|---|
| Digit V1 | 2019 | First humanoid version, upper body/arms added | 3 |
| Digit V2 | 2020-2021 | Last-mile delivery research collaboration with Ford (reported) | 3 |
| Digit (Commercial) | 2023 | Logistics-specialized design, improved End Effector | 6 |
| Digit (Next-gen) | 2024 | Battery expansion, payload increase | 7 |
Technical Specifications
Note: Specs below are collected from various sources and may vary by version. All units are standardized to SI (metric) with imperial units noted where needed.
Physical Specifications
| Item | Spec | Source |
|---|---|---|
| Height | 175 cm (5’9”) | 8 |
| Weight | 65 kg (143 lbs) | 8 |
| Walking Speed | 5 km/h | 8 |
| Payload | 16 kg (35 lbs), Next-gen: 23 kg (50 lbs) | 78 |
| Price | Approximately $250,000 (pilot program basis, reported) | 9 |
Degrees of Freedom (DoF)
| Version | DoF Configuration | Source |
|---|---|---|
| Initial | 16 DoF (legs 5x2 + arms 3x2) | 10 |
| Commercial | 28 DoF | 8 |
| Latest | 32 DoF (legs 6x2 + arms 7x2 + grippers 1x2 + waist 2 + neck 2) | 8 |
DoF configuration varies by version; figures may differ by source
Battery and Power
| Item | Spec | Source |
|---|---|---|
| Battery Type | Custom 1.2kWh Li-Po | 8 |
| Walking Time | 1.5 hours (initial) → 4 hours (2024 version) | 7 |
| Standby Time | 3 hours | 8 |
| Work:Charge Ratio | 4:1 (reported) | 7 |
| Auto Docking | Supported (added 2024) | 7 |
Sensors and Computing
| Item | Content | Source |
|---|---|---|
| Navigation | LiDAR | 8 |
| Depth Sensing | Intel RealSense x 4 | 8 |
| IMU | MEMS IMU | 8 |
| Proprioception | Absolute/Incremental encoders | 8 |
| Computing | Intel i7 dual CPU | 8 |
| Expansion Bay | Intel NUC / NVIDIA Jetson mountable | 8 |
End Effector (Gripper)
Digit adopts a modular End Effector approach 6:
- Suction Gripper: For flat surfaces of plastic totes/boxes
- MT 2 (Manipulation Tool 2): Newly introduced 2024
- Interchangeable design depending on task
Note: Compared to competitors (Tesla Optimus, Figure 02), dexterous hand technology is relatively simple; Agility focuses on logistics-specialized grippers
Amazon Partnership
Amazon invested in Agility Robotics through the Amazon Industrial Innovation Fund 11.
Deployment Stages
| Stage | Period | Content | Notes |
|---|---|---|---|
| Phase 1 | 2023-2024 | Pilot program (Sumner, WA, etc.) | Confirmed 11 |
| Phase 2 | 2025 | Expansion to 10+ logistics centers | Plan, reported 9 |
Initial Use Case
Tote Recycling: High-repetition task of collecting and moving empty totes after all inventory has been picked 11
Results (Reported)
| Metric | Value | Notes |
|---|---|---|
| Task Success Rate | 98% | After 18 months testing, estimate 9 |
| Hourly Cost | $10-12 | Compared to human worker $30, estimate 9 |
Note: Above figures are from third-party analysis reports, not Agility official announcements
GXO Deployment: Humanoid Commercialization Case in Logistics
Deployment Overview
| Item | Content |
|---|---|
| Location | Flowery Branch, Georgia (Atlanta suburbs) |
| Customer | SPANX omnichannel logistics center |
| Contract | Multi-year RaaS contract |
| Start | Late 2023 PoC → 2024 full deployment |
Source: 12
100K Tote Milestone (Fall 2025)
Digit demonstrated the following by achieving movement of over 100,000 totes 1:
- High-Volume Processing: Industrial-level throughput
- Reliability: Operation within live fulfillment workflows
- ROI Potential: Direction toward long-term ROI validation
Tasks Performed
- Picking up/putting down totes from Cobot AMRs
- Loading items onto conveyors
- Stacking containers at various floor positions
Orchestration
Manages Digit fleet through Agility Arc cloud platform, integrated with AMR vendors like MiR, Zebra Robotics 2.
RoboFab: Humanoid-Dedicated Manufacturing Facility
| Item | Content | Source |
|---|---|---|
| Location | Salem, Oregon (about 50km from engineering center) | 13 |
| Area | Approximately 6,500 m² (70,000 sq ft) | 13 |
| Opening | Late 2023 | 13 |
| Maximum Production Capacity | 10,000 units annually (target) | 13 |
| Employment Plan | 500 people (at max production) | 14 |
Production System Features
- Designed so each subassembly completes at the same time
- Simultaneous Assembly: Entire robot assembled simultaneously then moves to final test
- Scalable: Production increase by replicating each work cell
- High-rate manufacturing possible without line shutdown
Production Roadmap
| Period | Target |
|---|---|
| Year 1 | Hundreds of units (8 per shift) |
| Scale-up | Thousands of units (2026 target) |
| Maximum | 10,000 units annually |
Strategic Partnerships
| Partner | Period | Content | Source |
|---|---|---|---|
| Amazon | 2023 | Investment and warehouse deployment | 11 |
| GXO Logistics | 2023-2024 | Logistics sector RaaS contract | 12 |
| Zion Solutions | 2024.05 | Logistics/supply chain system integration | 15 |
| Ricoh USA | 2024.09 | Installation, service, customer support | 15 |
| MiR / Zebra | 2025 | AMR integration | 2 |
Competitive Positioning
Strengths
- Early Commercial Deployment Experience: Most advanced real operational experience in logistics sector
- Logistics Specialization: Focus on industrial rather than general-purpose
- Production Infrastructure: Mass production capability through RoboFab
Challenges
- Hand Technology: Simpler gripper compared to competitors (limitations when expanding beyond logistics)
- AI Capabilities: Room for advancement in End-to-End neural network approaches
- Versatility: Need to verify expandability beyond logistics areas
Competitor Comparison
| Company | Approach | Status |
|---|---|---|
| Agility | Logistics specialized, gradual expansion | Commercial operations |
| Tesla | General-purpose, mass production target | Pilot stage |
| Figure | High-performance AI (Helix VLA) | Testing |
| 1X | Home/general-purpose | Development |
Comparison is as of 2025, based on each company’s announcements and reports
Market Outlook
Labor Shortage Response
The logistics industry faces increasing volumes and persistent labor shortages, driving demand for humanoid automation solutions like Digit 16.
2025-2026 Outlook (Industry Estimates)
Note: Forecasts below are based on industry analysts and reports; actual results may differ
References
See Also
Footnotes
-
Digit 100K Tote Achievement Announcement - Agility Robotics Official ↩ ↩2
-
Livium Digit Specifications / QVIRO Digit Specifications ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14
-
Scaling Deep: Agility Robotics Analysis - Third-party analysis ↩ ↩2 ↩3 ↩4
-
RoboFab Official Page / RoboFab Opening Announcement ↩ ↩2 ↩3 ↩4
-
Contrary Research: Agility Robotics Partnership Analysis - Third-party analysis ↩ ↩2
-
Humanoid Robots 2025-2026 Outlook - Industry analysis ↩ ↩2 ↩3