The humanoid robotics market is moving into a more demanding phase. Technical novelty is no longer enough: the products that matter will be those that can perform useful work, integrate into existing operations, meet industrial safety requirements and deliver an economic return. Agility’s Digit offers a useful case study in how humanoid robotics is beginning to make that transition.

The market is beginning to ask a different question

For much of the recent humanoid robotics cycle, attention has naturally centred on technical demonstrations: a robot walking over uneven ground, handling an unfamiliar object or completing a task after a natural-language instruction. These achievements are important. They show how quickly the underlying hardware, controls and artificial intelligence are advancing. But a demonstration is not the same thing as a product.

A commercially useful industrial robot must do more than complete a task once. It must repeat that task reliably across thousands of cycles, operate safely around people, connect with existing equipment and software, recover from routine exceptions and produce a measurable benefit for the customer. It must also be deployable without requiring the customer to redesign an entire facility around it.

This distinction is becoming one of the most important filters in the sector. The long-term opportunity may not belong simply to the company that can build the most visually impressive humanoid. It may belong to the companies that can combine hardware, software, safety, service and operating data into a system that enterprises can deploy at scale.

Why the humanoid form factor has product value

Factories, warehouses and distribution centres were designed around people. Their aisles, shelves, workstations, containers and tools generally assume a worker with human reach, mobility and dexterity. Traditional automation can be extremely effective, but it is often fixed to a particular workcell or engineered for a narrow task. Expanding it may require new conveyors, cages, floor layouts or other infrastructure.

A humanoid form factor offers a different proposition: a mobile machine designed to work within human spaces and alongside existing automation. Digit, Agility’s bipedal mobile manipulation robot, is built at human scale to walk, lift and move materials through workflows that already exist. The objective is not to reproduce every human capability immediately. It is to address valuable, repetitive work while preserving the possibility of adding new tasks over time.

That flexibility matters economically. If one hardware platform can be redeployed across multiple workflows as its software capabilities improve, the customer may gain more utility from the same installed asset. The humanoid’s potential advantage is therefore not its appearance; it is compatibility with the physical world businesses have already built.

Useful work is a stronger benchmark than a demonstration

Digit’s deployment with GXO provides a concrete example. Following a proof-of-concept pilot, GXO and Agility entered a multi-year Robots-as-a-Service agreement in 2024 for Digit robots to operate in a live logistics environment. At the SPANX facility in Georgia, Digit works with autonomous mobile robots, moving totes from those systems onto conveyors as part of the fulfilment workflow.

By November 2025, Agility reported that Digit had moved more than 100,000 totes at the facility. The number is meaningful not because tote handling represents the final ambition for humanoids, but because it tests the less glamorous attributes that commercial products require: repeatability, uptime, safe interaction, integration and sustained throughput.

This is the bridge from laboratory capability to customer value. Early industrial use cases are likely to be bounded and repetitive, particularly where labour is difficult to recruit or where tasks are physically demanding and ergonomically unattractive. Success in those workflows can create operating data, customer trust and deployment experience that support a broader set of capabilities later.

The product is the system, not only the robot

It is tempting to evaluate humanoid companies almost entirely through their hardware. In practice, the robot is only one layer of the commercial product. A customer also needs a way to map the facility, define workflows, coordinate a fleet, monitor performance, troubleshoot problems and integrate the robots with warehouse systems and other machines.

Agility addresses this through Arc, its cloud automation platform. Arc is designed to manage the deployment lifecycle and orchestrate Digit alongside equipment such as autonomous mobile robots. This software layer is strategically important: it turns individual machines into an operational fleet and gives customers a practical interface for managing work rather than managing robotics research.

Over time, this systems layer may also become an important source of differentiation. Hardware will continue to improve across the industry, but deployment knowledge, workflow integrations, safety processes and real-world performance data can compound. A robot that is technically capable but difficult to integrate may create less value than a slightly less spectacular machine supported by a mature operating platform.

Safety, reliability and serviceability are core features

Industrial customers do not treat safety as an optional add-on. Robots working near employees must satisfy rigorous risk assessments, site requirements and recognised standards. In 2025, Agility said Digit passed an OSHA-recognised Nationally Recognized Testing Laboratory field evaluation at an ecommerce fulfilment site. The evaluation was site-specific, but it demonstrated a repeatable route toward compliant deployment in a live workplace.

Reliability and serviceability matter for the same reason. Every hour of unavailable capacity weakens the customer’s return, while difficult maintenance raises the cost of scaling a fleet. Features such as autonomous charging, fleet monitoring and systematic fault recovery may attract less attention than new AI behaviours, but they are fundamental to whether robotics can function as dependable infrastructure.

The Robots-as-a-Service model can further reduce adoption friction. Instead of asking customers to make a large upfront equipment purchase and absorb all technology risk, the supplier can package hardware, software and support as an ongoing service. This aligns the commercial proposition more closely with productive use, although it also places greater responsibility on the provider to finance, maintain and continually improve the installed fleet.

From one task to a broader platform

The near-term case for humanoids is based on specific industrial jobs; the larger opportunity depends on expanding what the same platform can do. Advances in Physical AI are helping robots learn behaviours from demonstrations, simulation and real-world experience rather than relying only on rigid, manually programmed instructions.

Agility’s development strategy reflects that progression. In July 2026, the company announced a 60,000-square-foot Fremont facility focused on training, testing and advancing the AI systems that enable Digit to learn new skills. The site complements RoboFab, its manufacturing operation in Oregon. Together, those investments illustrate the two-sided challenge of productisation: the company must improve both intelligence and the ability to manufacture and support physical machines.

The expansion path should still be viewed with discipline. A robot that can execute one validated workflow does not automatically become a general-purpose worker. New tasks introduce different objects, failure modes, safety considerations and integration requirements. The credible route is likely to be incremental: establish reliability in a constrained application, add adjacent skills, and broaden autonomy as the operating evidence grows.

Commercial signals are beginning to emerge

Beyond the GXO deployment, Agility has announced commercial agreements with customers including Toyota Motor Manufacturing Canada and Mercado Libre. In connection with its proposed public-market transaction announced in 2026, the company also stated that it had secured more than $300 million of multi-year Digit v5 orders, subject to the achievement of certain contractual milestones, and had a pipeline of more than 30 customers.

These figures should be interpreted carefully: contracted orders subject to milestones are not the same as recognised revenue, and a customer pipeline is not a guarantee of deployment. Nevertheless, they indicate that enterprise interest is progressing beyond experimentation. The next test will be conversion—whether orders translate into installed fleets, recurring utilisation and attractive unit economics.

What investors should watch

• Deployment velocity: How quickly pilots and orders become active commercial fleets.
• Utilisation and uptime: Whether robots remain productive for enough hours to support compelling customer economics.
• Workflow expansion: Whether the same platform can add adjacent tasks without extensive hardware redesign.
• Safety at scale: Whether compliance and risk-assessment processes can be repeated efficiently across sites.
• Manufacturing economics: Whether production volume lowers cost while preserving quality and reliability.
• Software and data advantages: Whether fleet orchestration, integrations and deployment data create durable differentiation.

Hillside’s perspective

The humanoid robotics market is still early, and meaningful technical and commercial risk remains. Hardware must become more capable and economical; autonomy must improve without compromising safety; and suppliers must prove that they can manufacture, deploy and support fleets across multiple customer environments.

Yet the basis of competition is becoming clearer. The sector is moving from a race to demonstrate possibility toward a race to deliver repeatable value. In that environment, commercial deployments, operating data, safety credentials, software integration and customer economics matter at least as much as the latest demonstration.

For Hillside Enterprises, our investment in Agility provides exposure to this transition from humanoid robotics as an emerging technology to humanoid robotics as a deployable product. Digit’s significance lies not only in what the robot can do today, but in the product architecture being built around it: a human-compatible form factor, a fleet-management layer, a service model, real-world operating evidence and a roadmap for adding capability over time.

The ultimate winners in humanoid robotics will not be determined by novelty alone. They will be determined by which companies can make advanced machines useful, safe, reliable and economically valuable in the environments where work already happens.