Chinese humanoid robots are reshaping warehouse operations across Asia, but the real business opportunity isn't in building the machines,it's in making them work reliably at scale. As manufacturers like Unitree ship robots to Japan and beyond, local service partners are emerging as the critical link between cutting-edge hardware and practical, profitable deployment. This shift reveals a fundamental truth about the robotics industry: shipping a robot is easy; making it solve real problems is the hard part.

How Are Humanoid Robots Actually Getting Deployed in Real Warehouses?

The journey from robot prototype to working warehouse asset involves far more than unboxing and powering on. When Agility Robotics' Digit humanoid began operating in a GXO logistics facility in late 2023, the deployment required careful integration with existing systems. By November 2025, those robots had moved more than 100,000 totes at the facility, working alongside autonomous mobile robots and conveyors. The success came from a multi-year service agreement that treated the robots not as one-time equipment purchases, but as ongoing operational assets managed through a structured partnership.

Similarly, at Tokyo's Haneda Airport, Chinese-made humanoids from GMO AI & Robotics are undergoing a three-year trial that began in May 2026 and is scheduled to run through 2028. Before these machines could handle baggage and cargo in the airport's tight spaces, JAL Ground Service had to map which tasks were suitable and what safety conditions the robots needed to meet. The machines arrived with the ability to walk, carry, and balance, but their specific airport jobs still had to be programmed into software. This gap between general capability and specific application is where the real business opportunity emerges.

Why Are Local Service Partners Becoming More Valuable Than Robot Manufacturers?

The economics of robot deployment are shifting in unexpected ways. When Unitree Robotics appointed GMO as its authorized distributor in Japan in June 2026, the agreement covered far more than just sales. GMO now handles implementation support, software development, secure communications, maintenance, and operations. The company is also renting humanoids for proof-of-concept trials, allowing customers to test robots without committing to a purchase and giving GMO hands-on experience across different sites.

This model works because local partners can solve problems that manufacturers cannot practically handle from their headquarters. A routine fault for Unitree's engineers in Hangzhou may be unfamiliar to the technician standing beside the machine in Tokyo. JAL needs to know who will restore service and explain what happened. When workflows change, the motion program may need to change with it. Each new site brings another round of integration work. GMO is positioning itself to earn revenue not from a single hardware sale, but from rentals, software customization, maintenance, and ongoing operations after installation.

What Are the Key Steps for Scaling Humanoid Deployment in New Markets?

• Workflow Mapping: Before deployment, customers and service partners must identify which specific tasks are suitable for humanoid robots, what safety conditions must be met, and how the machines fit into existing operations without requiring expensive facility modifications.
• Software Customization: General-purpose robots arrive with basic capabilities like walking and carrying, but their specific job functions must be programmed into motion software tailored to each customer's environment and workflow requirements.
• Local Service Infrastructure: Establishing maintenance, spare parts availability, and technical support in the customer's region is essential; relying on manufacturer support across time zones creates unacceptable delays when baggage or packages are waiting to be moved.
• Escalation Pathways: Clear agreements must define what problems can be solved locally, when manufacturer engineers become involved, and who remains responsible during the resolution process to avoid confusion and service gaps.
• Pilot-to-Production Transition: Extended trials like Haneda's three-year program allow for repeated adjustment and extra supervision, but commercial operations cannot tolerate the same level of downtime and rework.

The practical test of this model will come when GMO takes what it learns at Haneda to another airport without rebuilding the entire project from scratch. If the company can replicate its success across multiple sites, it will have created something larger than a simple distribution business. As one analyst noted, "The machines are Chinese. The business that makes them useful may be built in Japan".

This dynamic mirrors what happened with Agility Robotics' Digit in the United States. GXO made the transition from testing to commercial operation by signing a multi-year Robots-as-a-Service agreement with Agility, which supplied the robots, fleet-management platform, and much of the deployment expertise. Agility's software covered facility mapping, workflow definition, fleet operations, and troubleshooting. At Haneda, comparable functions are divided among Unitree, GMO, and JAL rather than held by one supplier, creating a more distributed responsibility model.

What Challenges Could Derail This New Business Model?

The success of local service partners depends on having genuine technical control and independence from the manufacturer. If Unitree retains tight control over software tools and updates, GMO could be answerable to JAL without being able to resolve serious faults on its own. A response-time promise is only as good as the access behind it. Lower-priced hardware does not settle the economics either; extensive customization can erase the initial savings, while a local service agreement may still leave the customer dependent on the manufacturer's headquarters for critical decisions.

The second contract GMO signs will be more revealing than the Haneda demonstration. If the company can take what it learns there to another airport without rebuilding the project from scratch, it will have proven that the model works at scale. If each deployment requires starting over, then the business remains a high-touch, low-margin operation that cannot compete with manufacturers' own direct sales teams.

For customers like JAL, the key question is whether GMO's technicians receive the training, spare parts, and software access to handle routine failures without waiting for Unitree engineers in China. Clear escalation paths and defined responsibilities are essential to avoid situations where a customer is left waiting while a fault is referred across time zones and neither party takes full responsibility for the delay.

The robotics industry is learning that the future belongs not to the companies that build the best robots, but to the companies that make those robots reliably useful in real-world operations. For Agility Robotics and Unitree, that means finding partners who can bridge the gap between manufacturing capability and customer success. For companies like GMO, it means building a sustainable business around solving the problems that manufacturers cannot practically solve from afar.

" }