Quick summary: Agility Robotics is the Oregon-based humanoid robotics company that has the most-deployed-in-production humanoid in the world. Their robot Digit was the first humanoid to be deployed at a commercial customer site (GXO Logistics, June 2024). As of late 2025, Digit has moved over 100,000 totes at GXO’s Flowery Branch / Spanx facility, GXO has signed an industry-first multi-year Robots-as-a-Service agreement with Agility, and Toyota Motor Manufacturing Canada has signed a post-pilot deployment agreement. CEO is Peggy Johnson (former Magic Leap CEO, ex-Microsoft EVP). This guide covers what Digit does, who’s deploying it, how Agility’s market position is unusual, and how it compares to Figure, Boston Dynamics, and Tesla. Written for beginners. Updated 2026-05-15.
If you read humanoid-robotics news only through the lens of “who raised the most money,” you might miss Agility Robotics entirely. Figure has raised $1.9 billion. Apptronik is at $1 billion. Tesla’s parent company is worth a trillion. Agility has raised roughly $300 million — an order of magnitude less than Figure. And yet Agility’s robot Digit is the one that’s actually been working in a paying customer’s facility longer than any other humanoid in the world, the only one that’s moved over 100,000 totes in real commercial operation, and the first to sign a multi-year Robots-as-a-Service contract with an enterprise logistics customer.
This is the explainer. What Agility does differently, why “less money but more deployment” can be a viable strategy, and where the company fits in a humanoid market dominated by louder competitors.
Get Smarter About AI Every Morning
Free daily newsletter — one story, one tool, one tip. Plain English, no jargon.
Free forever. Unsubscribe anytime.
What is Agility Robotics in plain English?
Agility Robotics is an Albany, Oregon-based humanoid robotics company that makes Digit, a bipedal robot designed for warehouse and logistics work. Where Figure and Tesla make humanoids with arms, hands, legs, and a torso, Digit takes a deliberately different design path: a bipedal lower body (no toes — its feet end in a single point and the knees bend backwards like an ostrich’s), simpler arms, and an unfussy head. The robot is purpose-built for moving boxes and totes — not for assembling cars, not for picking up coffee mugs, not for dancing on stage. The narrow focus is the point.
The company was founded in 2015 by Damion Shelton and Jonathan Hurst as a spinoff from Oregon State University’s robotics research. Hurst’s academic work on dynamic-balance bipedal locomotion was the technical seed of what became Digit. Peggy Johnson — formerly CEO of Magic Leap and EVP of Business Development at Microsoft — took over as CEO in 2023, bringing enterprise-sales muscle to a company that had focused primarily on engineering.
What does Digit actually do?
Digit is built to do one specific job extraordinarily well: take totes off a shelf or conveyor, walk them somewhere else, and put them down. In warehouse operations this single task — called “tote handling” — is one of the most physically demanding and repetitive in the building. Humans doing it all day burn out fast. Traditional industrial automation (conveyor systems, automated guided vehicles) can handle the long-haul movement, but the “pick up the tote, carry it across uneven floor, place it on a tugger” segment is what kept human workers tethered to the task.
| Spec | Detail |
|---|---|
| Height | ~5’9″ |
| Weight | ~141 lb |
| Payload | 35 lb (designed for typical warehouse totes) |
| Reach | ~6 ft |
| Speed | Up to 3.3 mph |
| Operation time | ~16 hours per shift; battery hot-swap supports continuous |
| Leg design | Reverse-knee bipedal (ostrich-style); deliberately non-anthropomorphic |
| Brain | Agility-developed control + planning stack |
The reverse-knee design isn’t a styling choice. It’s an engineering decision: the bird-leg topology is mechanically efficient for the specific task of carrying a load while walking, and avoids many of the balance edge cases that human-style knees introduce. Agility’s bet was that for warehouse work, looking like a human is less important than working like an efficient quadruped that happens to walk on two legs.
Where is Digit actually deployed?
| Customer | Facility | What Digit does | Status |
|---|---|---|---|
| GXO Logistics | Flowery Branch, GA (Spanx fulfillment) | Tote handling alongside cobots | 100,000+ totes moved (fall 2025); industry-first multi-year RaaS contract |
| Toyota Motor Manufacturing Canada | Ontario plant | Loading/unloading totes from automated tuggers | Post-pilot deployment agreement signed; 7 more units coming online |
| Amazon | Multiple facilities | Tote handling, internal logistics | Multi-site pilot deployment |
| Spanx | Fulfillment operations (via GXO) | Same operational base as above | Production deployment |
What’s noteworthy about this list is that GXO’s deployment is the first publicly-disclosed humanoid that’s working under a Robots-as-a-Service contract — meaning GXO doesn’t buy Digits outright, they pay Agility for the work the robots do. The pricing is per robot per hour (industry-reported around the $10-30/hour range, with the actual GXO rate not disclosed publicly). This is structurally similar to how AWS sells cloud computing: the customer pays for usage, not for ownership.
The Toyota deal is the most-watched recent deployment because it’s Agility’s first formal commitment in automotive manufacturing, the same vertical where Figure has BMW, Boston Dynamics has Hyundai, and Apptronik has Mercedes. The seven additional Digits being deployed at Toyota’s Canadian plant are an explicit “we want more if these work” signal from one of the world’s largest automakers.
Why does Agility’s strategy work despite having less capital?
Three things. First, the deliberate narrow-task focus. While Figure is trying to do everything from automotive assembly to eventual home use, Agility focuses entirely on warehouse and tote-handling work. A robot designed for one task is cheaper to engineer, faster to deploy, and easier to validate than a robot designed for general-purpose use.
Second, the non-humanoid form factor. Reverse-knee bipedal locomotion is mechanically simpler than human-style locomotion. Agility doesn’t have to solve the harder problems (full-range arms, dexterous hands, human-shape pelvis dynamics) that drag down everyone else’s engineering timelines. The robot looks less impressive in promotional videos and works more reliably in production.
Third, the Robots-as-a-Service business model. Selling robots outright to enterprise customers is hard — long procurement cycles, large capex commitments, customer reluctance to make multi-year hardware bets. Selling robots as a service moves the conversation to operational cost (which warehouses already understand) and avoids the asset-on-balance-sheet conversation entirely.
Combined: Agility produces a lower-cost, narrower-purpose robot with a customer-friendly buying model, and as a result has more total humanoid-hours in production than any other company on the planet.
Who has invested in Agility Robotics?
Agility’s investor list is unusual in that strategic operators dominate over financial-only firms. Amazon’s Industrial Innovation Fund made an early growth investment, signaling Amazon’s interest in Digit for its own warehouses (and giving Amazon a strategic option on the humanoid logistics category). DCVC and Playground Global led on the venture side. GMO Internet Group, Sumitomo Mitsui Banking, MFV Partners, and Manifest Growth Partners have also participated. Total raised is in the $300 million range; the company was last reported at a billion-dollar-plus valuation after its 2024 Series C.
This investor pattern is structurally different from Figure’s (where the biggest names are Microsoft, NVIDIA, OpenAI Startup Fund). Agility looks like a company optimizing for being a reliable supplier to large industrial customers. Figure looks like a company optimizing for being a platform with strategic AI partnerships. Both are defensible. Both could win.
How does Digit compare to Atlas, Figure 03, and Optimus?
| Robot | Form | Payload | Production-deployed hours | Pricing model |
|---|---|---|---|---|
| Digit | Reverse-knee bipedal; warehouse-focused | 35 lb | Most of any commercial humanoid (GXO, Toyota, Amazon) | Robots-as-a-Service |
| Figure 03 | Anthropomorphic humanoid; general-purpose | 20 kg (~44 lb) | BMW Spartanburg deployment (Figure 02) | Enterprise contracts (terms not public) |
| Atlas (electric) | Anthropomorphic; high-DOF | 50 kg (110 lb) | Hyundai RMAC + Google DeepMind (start March 2026) | Enterprise contracts |
| Optimus V3 | Anthropomorphic; manufacturing-focused | Not specified publicly | Internal Tesla only as of mid-2026 | Not yet sold externally |
| Apptronik Apollo | Anthropomorphic; industrial | ~25 kg | Mercedes pilot, GXO R&D | Enterprise contracts |
Digit wins on real-world deployment hours and on customer count. Atlas wins on technical depth and payload. Figure 03 wins on marketing breadth and capital raised. Optimus wins on potential scale. Apollo wins on automotive-OEM relationships. None of these is the “best” humanoid in a general sense; they’re each optimized for different things. The 2030 picture probably has Digit dominant in warehouse logistics, Atlas/Figure/Optimus competing in heavy manufacturing, and Apollo holding a strong position in European automotive.
What can go wrong for Agility?
- Capital disadvantage. If the humanoid race becomes about raw scale of investment in AI training, manufacturing capacity, and global expansion, Figure and Tesla will out-spend Agility ten-to-one.
- The “narrow task” advantage becomes a “narrow market” disadvantage. If general-purpose humanoids reach a price-and-capability point where they can also do warehouse tote-handling, the specialized advantage of Digit evaporates.
- Amazon goes its own way. Amazon has invested in Agility and is a deployment customer, but Amazon also operates its own robotics R&D and is reportedly developing alternative humanoid platforms. If Amazon shifts to its own hardware, Agility loses a major strategic position.
- RaaS pricing pressure. The hourly economics of Robots-as-a-Service depend on robot uptime, customer demand, and the cost of competing solutions (including human labor). All three move against Agility if competitors crack the same model with lower per-hour pricing.
- Talent retention. Younger humanoid companies pay aggressively in equity. Agility’s longer-tenured engineering team is a target for poaching from better-funded competitors.
- Form-factor lock-in. If the market consolidates around anthropomorphic humanoids for reasons that have less to do with task efficiency and more to do with general-purpose flexibility (or just consumer aesthetics), Digit’s reverse-knee design could become a stranded technical bet.
Learn Our Proven AI Frameworks
Beginners in AI created 6 branded frameworks to help you master AI: STACK for prompting, BUILD for business, ADAPT for learning, THINK for decisions, CRAFT for content, and CRON for automation.
Frequently asked questions
Can my company buy a Digit?
Yes, but through enterprise procurement, not as a catalog purchase. Agility’s primary commercial model is Robots-as-a-Service, where the customer pays a per-hour or per-robot operational fee. Outright unit purchases have been quoted in industry reporting at low six figures. For a small or mid-size warehouse, the RaaS path is almost always the right one.
Why does Digit have ostrich-style knees?
It’s mechanically more efficient for the specific task of walking while carrying a load. The reverse-knee design lowers the robot’s center of mass during locomotion, simplifies the dynamics of weight transfer between steps, and avoids many of the balance edge cases that human-style knees introduce. The form factor is an engineering decision, not aesthetic.
Who is the CEO of Agility Robotics?
Peggy Johnson, who became CEO in 2023. Johnson previously led Magic Leap and spent 24 years at Microsoft, ultimately as EVP of Business Development. Her appointment signaled Agility’s strategic shift from research-led to enterprise-sales-led, which has paid off in the GXO, Toyota, and Amazon deployments.
Does Digit have arms and hands?
Yes, but they’re deliberately simpler than the arms on a Figure or Tesla humanoid. Digit’s arms are designed to grasp and lift totes — not to manipulate fine objects, use hand tools, or perform delicate assembly. The simpler arms are part of why Agility ships at higher reliability for the narrow task they target.
How does Robots-as-a-Service pricing work?
The customer pays an hourly or per-robot monthly fee for the work the robots do. Agility maintains the robots, provides software updates, and handles repairs. The customer gets robot labor without buying the hardware. Per-hour pricing in industry reporting has been quoted in the $10-30 range, though the GXO contract terms specifically are not public.
Is Agility Robotics publicly traded?
No. Agility is a private company. With the Toyota and GXO contracts now public, IPO speculation has increased — but no formal announcement has been made as of mid-2026.
Sources
- Agility Robotics — Digit product information and deployments
- Agility Robotics — Digit Moves Over 100,000 Totes in Commercial Deployment
- Agility Robotics — GXO Multi-Year Agreement Announcement
- GXO Logistics — Industry-first agreement with Agility
- The Robot Report — Toyota Motor Manufacturing Canada deployment
- Humanoids Daily — Sector news