Agility Robotics & Digit — Complete Profile & Analysis 2026

Company Overview

Agility Robotics occupies a unique position in the humanoid robotics landscape: it is the only company with a humanoid robot actively deploying in Amazon fulfillment centers. This gives Agility something no other humanoid company has — real-world operational data from one of the most demanding logistics environments on the planet.

Founded in 2015 as a spin-off from Oregon State University's AMBER Lab, Agility has deep academic roots in bipedal locomotion research. The company's first product, Cassie, was a legs-only bipedal platform that was sold to more than 30 research institutions and set a Guinness World Record for the fastest 100-meter dash by a bipedal robot. The locomotion control expertise developed through Cassie directly informed Digit — Agility's full humanoid, which adds an upper body, arms, and manipulation capabilities to Cassie's proven bipedal base.

Amazon's majority investment in 2023 was the pivotal moment in Agility's trajectory. Rather than competing across every possible humanoid use case, the Amazon relationship allows Agility to focus deeply on warehouse logistics — a massive, well-defined, and economically motivated market. The trade-off is strategic narrowing: Digit is optimized for tote handling and warehouse navigation, not for manufacturing manipulation or outdoor tasks. But this focus may be exactly the right strategy. Finding one use case that works at scale is more valuable than being mediocre at ten.

Digit: Product Deep-Dive

Source: Agility Robotics disclosures, SVRC Research

Warehouse-Optimized Design

Digit's design reflects a deliberate choice: optimize for warehouse logistics rather than general-purpose capability. At 175 cm and 65 kg, Digit fits in standard warehouse aisles, reaches standard shelving heights, and can handle standard totes (the plastic bins that move goods through Amazon fulfillment centers). The arms are designed for tote grasping and placement rather than fine manipulation — a practical decision that reduces mechanical complexity while addressing the primary use case.

This warehouse-first design philosophy extends to the perception system, which is optimized for structured indoor environments with known shelf layouts and tote geometries, and the navigation system, which handles the specific challenges of warehouse floors (smooth concrete, narrow aisles, dynamic foot traffic from human workers and AMRs).

Bipedal Advantage in Warehouses

The case for a bipedal humanoid in warehouses (as opposed to wheeled robots or fixed arms) rests on a specific argument: warehouses are designed for humans, so a human-form-factor robot can operate in existing facilities without modification. Digit can walk up stairs, step over obstacles, and navigate between shelving units in ways that wheeled robots cannot. In Amazon's massive fulfillment centers — some exceeding 1 million square feet — this versatility translates into significant deployment flexibility.

The Amazon Connection

Amazon's majority investment in Agility Robotics is the most significant strategic move in humanoid robotics deployment. Amazon operates more than 1,000 fulfillment centers globally, employs more than 1.5 million warehouse workers, and has invested tens of billions in logistics automation (including the $1.7B acquisition of Kiva Systems/Amazon Robotics in 2012). Adding humanoid robots to this infrastructure is a logical next step.

Deployment Scale

Agility has been deploying Digit in Amazon fulfillment centers since late 2023, with the number of units and facilities growing through 2024-2026. While exact deployment numbers are not publicly disclosed, the program has expanded from initial pilot facilities to multiple fulfillment centers across the US. The deployment focuses on tote handling — moving plastic bins between shelves, conveyors, and workstations — a task that is repetitive, physically demanding for human workers, and well-suited to automation.

Source: SVRC Research estimates based on public reporting

Strategic Implications

The Amazon relationship creates both an opportunity and a constraint. The opportunity is massive: if Digit proves ROI at Amazon scale, the addressable market within Amazon alone could be thousands of units. The constraint is strategic narrowing: deep Amazon integration may make it difficult for Agility to serve other customers who compete with Amazon or have different operational requirements. This is a classic platform-company dynamic — and one that Agility will need to navigate carefully as it matures.

Technology Deep-Dive

AMBER Lab Heritage

Agility's bipedal locomotion technology traces directly to the AMBER (Advanced Mechanical Bipedal Experimental Robotics) Lab at Oregon State University, led by Dr. Jonathan Hurst (Agility's co-founder and CTO). The AMBER Lab pioneered a mathematical approach to bipedal walking based on Hybrid Zero Dynamics (HZD) — a control framework that models the alternation between contact and flight phases in bipedal locomotion as a hybrid dynamical system. This theoretical foundation gives Agility's control algorithms a level of formal guarantees that pure learning-based approaches do not provide.

Locomotion Control

Digit's walking controller combines the classical HZD framework with modern reinforcement learning policies. The classical controller provides stability guarantees and efficient locomotion on flat surfaces, while RL policies handle terrain adaptation, obstacle avoidance, and recovery from unexpected disturbances. This hybrid approach — classical for baseline stability, learned for adaptation — is similar to what Boston Dynamics has adopted, though implemented differently.

ROS Integration

Digit uses ROS (Robot Operating System) for its software stack, which enables integration with the broader robotics software ecosystem. For Amazon's deployment, this ROS-based architecture connects Digit to Amazon's existing warehouse management and fleet coordination systems, allowing Digit to receive task assignments, navigate to specified locations, and report task completion through standard interfaces.

Source: SVRC Research assessment

Key Milestones

Market Position & Competitive Landscape

Competitive Strengths

• Amazon deployment data: More real-world humanoid operational hours than any competitor. This data compounds — every hour of operation improves Digit's AI models.
• Academic foundations: OSU AMBER Lab heritage provides deep theoretical understanding of bipedal dynamics that pure-learning approaches lack.
• Vertical focus: Warehouse-first strategy enables deep optimization rather than spreading thin across use cases.
• RoboFab: Dedicated manufacturing facility in Salem gives Agility production control that leased-facility competitors lack.
• Amazon ecosystem: Integration with Amazon's warehouse management systems provides a deployment pathway that other humanoid companies must build from scratch.

Competitive Vulnerabilities

• Amazon dependency: Majority Amazon ownership and warehouse focus could limit commercial appeal to non-Amazon customers.
• Narrow versatility: Warehouse optimization means Digit is less capable than general-purpose humanoids for manufacturing, outdoor, or manipulation-heavy tasks.
• Funding gap: ~$150M known funding is modest compared to Figure ($675M) and Pi ($400M). Amazon provides strategic capital but public fundraising has been limited.
• Pricing pressure: Unitree G1 at $16K and H1 at $90K apply pressure even if Digit's warehouse-specific capabilities are superior.

SVRC's Assessment

Agility Robotics represents the "pick one use case and go deep" strategy for humanoid commercialization, and there is a strong argument that this is the correct approach. The humanoid robotics industry is littered with companies trying to build general-purpose robots that end up being mediocre at everything. By focusing exclusively on warehouse logistics — and having the world's largest warehouse operator as both customer and investor — Agility has a clearer path to commercial viability than most of its competitors.

The data advantage is real and compounding. Every hour that Digit operates in an Amazon fulfillment center generates tote manipulation data, navigation data, and failure-recovery data that feeds directly into improving the next generation of control models. No university lab, no controlled test environment, and no simulation can replicate the diversity and scale of real Amazon warehouse conditions. This data moat will widen over time.

The strategic risk is Amazon dependency. If Agility becomes too tightly integrated with Amazon's specific requirements, it may struggle to serve other warehouse operators (who may have different tote sizes, shelf configurations, or operational workflows) or to expand into non-warehouse applications. The best outcome for Agility is to use the Amazon deployment as a proving ground for technology that eventually serves the broader logistics industry — but that pivot requires deliberate strategic planning.

For buyers in warehouse logistics: Digit is the most deployment-ready humanoid for tote handling and warehouse navigation available today. For general-purpose humanoid research: evaluate Figure 02, Unitree H1, or Unitree G1 depending on your budget and requirements.

Frequently Asked Questions