Editor’s note: This article is republished from Jumpstart, a TechNode Group publication.

A warehouse can look full while much of its space remains empty.

The shelves hold inventory, the aisles allow workers and equipment to move, and the space above them waits for a practical way to be used. Storing goods higher is easy enough. Retrieving the right box quickly, safely and repeatedly is the harder problem.

Hai Robotics has built its business around that distinction.

The Shenzhen company makes robots that retrieve individual containers from warehouse racks and bring them to workstations. Its newer machines climb the racks themselves, turning height into storage capacity without requiring workers to follow inventory upward.

It is a striking image: a compact robot ascending a wall of boxes. But the more consequential change happens around the machine. Aisles can become narrower. Inventory can be arranged differently. Workers can remain at picking stations while goods travel toward them. Software determines which container should move, which robot should retrieve it and which order should be completed next.

The warehouse begins to function as a coordinated system rather than a building through which people carry things.

Founded in 2016, Hai Robotics is now seeking a Hong Kong listing. It refreshed its application on September 13, with Goldman Sachs and CITIC Securities as joint sponsors. Its revenue reached RMB2.02 billion in 2025, while sales outside mainland China accounted for just over half of revenue in the first half of 2026. According to research commissioned for the filing, it was the world’s largest provider of autonomous case-handling robot solutions by revenue and shipments in 2025. That ranking refers to a specific segment of warehouse automation, rather than the entire robotics industry. Latest application overview

The listing brings a less visible part of China’s robotics expansion into focus. Hai’s machines do not need to resemble people or perform theatrical demonstrations. Their value is measured in containers retrieved, orders completed and space recovered.

For Hai itself, the next challenge is equally practical: turning successful warehouse projects into a profitable global business.

Start With the Box

Warehouse automation existed long before Hai Robotics.

Conveyors, cranes and automated storage systems have moved goods through industrial facilities for decades. More recently, mobile robots began transporting shelves or containers across warehouse floors. Each approach reflects a different answer to the same question: how should inventory reach the person or machine that needs it?

Hai concentrated on the individual case.

A case might be a plastic tote containing small products or a carton holding stock. Retrieving that container allows an operator to select the required item without walking through rows of shelving to find it.

The industry calls this “goods-to-person” automation. The phrase describes a reversal in the usual relationship between workers and inventory. Instead of sending a worker to a product, the system sends the product to a worker.

For a warehouse handling thousands of distinct items, that reversal changes more than travel time. It can make the picking process easier to organize, reduce repeated lifting and allow inventory to be stored in locations chosen for density rather than convenient human access.

Hai’s early development followed this logic. The company introduced a single-case robot system in 2017 and a multi-case robot in 2018. The latter could retrieve and carry several containers during one journey, improving the usefulness of each trip through the warehouse.

Founder and chief executive Chen Yuqi, known internationally as Richie Chen, described the early design philosophy in terms of affordability and ease of use. The ambition was to make automation practical for warehouses and factories that needed to improve existing operations. Hai’s early company account

That constraint would become central to the company’s identity. A robot that works only inside an entirely rebuilt facility has a different market from one that can be fitted into a functioning warehouse.

Building Around the Existing Warehouse

A warehouse is an awkward place to introduce new technology.

Its floor may be uneven. Its shelving may have been purchased years earlier. Its inventory changes with customers and seasons. Its operators cannot necessarily suspend order fulfillment while engineers install a new system.

An automation supplier therefore has to solve two problems at once. It must improve the operation, and it must make the transition tolerable.

Hai’s approach combines robots, storage racks, workstations and software into a system that can be configured for different facilities. The company develops the machines, but the commercial product is the complete flow of goods.

That distinction explains why a warehouse robotics business can be more complicated than its hardware suggests. A customer buys an operating result. Delivering it requires the robots to communicate with order systems, identify containers, coordinate movement and recover when something goes wrong.

A fast robot can still spend much of its time waiting for a congested workstation. A dense storage layout can become inefficient if frequently requested products sit in the wrong locations. Adding machines can increase throughput—or simply create more traffic.

HaiQ, the company’s software platform, coordinates tasks, inventory movements and robot assignments. It is the layer that turns a collection of machines into an automated warehouse.

For customers, the important question is how the whole system performs during an ordinary shift and a seasonal peak. The answer depends on engineering and deployment as much as on the robot’s specifications.

Learning to Climb

Hai’s latest major product step takes the original case-handling idea into the vertical dimension.

Introduced in 2025, HaiPick Climb uses compact robots that move along guide channels attached to warehouse racks. They ascend to retrieve containers and return them to the flow feeding picking stations.

Attaching the climbing system to one side of the rack is an important part of its design. Hai says this reduces infrastructure requirements and allows greater tolerance for variations in racks and floors, making installation easier in existing facilities. Product launch

The commercial logic is straightforward. Warehouse operators pay for a building’s footprint, but increasingly need to store more products and process more orders within it. Vertical retrieval makes more of the building usable. Narrower aisles can release additional space.

The gains remain specific to each installation. Inventory mix, building height, order patterns and workstation capacity all affect the result. A headline storage-density figure cannot describe every warehouse.

Nevertheless, the direction is clear. Hai is making the building’s geometry part of its product.

The company reported in August 2026 that more than 10,000 HaiClimber robots had been contracted for customer projects across 12 countries since the system’s global introduction. “Contracted” is an order milestone, rather than a count of robots already operating. Named customers included Panasonic, Arvato, ITOCHU and ANTA. Hai also reported that eight customers had expanded an existing system or selected the product again for another project. Company announcement

Repeat purchases matter in this industry. A warehouse becomes a demanding reference site once its daily operations depend on the system. A customer willing to extend that dependence provides stronger evidence than a demonstration.

A Global Business Built One Site at a Time

Hai’s expansion has increasingly taken it beyond China.

Revenue from outside mainland China rose from 24.2% of the total in 2023 to 47% in 2025 and 50.2% in the first half of 2026. Overseas orders represented more than 55% of orders during that latest period. Application overview

This is a substantial change in the company’s business. International operations have become central to its growth, bringing Hai into facilities with different labor costs, property constraints and expectations for technical support.

The company serves ecommerce, logistics, apparel, manufacturing and other industries. Its global network includes regional offices and implementation partners, reflecting the local work required to install and maintain warehouse systems. Company profile

Selling a robot overseas is only the beginning. A supplier must understand the facility, integrate with existing software, train operators and provide maintenance. It may need spare parts and engineers close enough to respond when a system stops.

That makes overseas expansion both attractive and expensive.

A successful project can create a relationship that lasts through several warehouse expansions. It can also become a reference for other customers in the region. But building the sales and service organization to win those projects requires spending before revenue is recognized.

Hai’s global ambition therefore depends on how consistently it can repeat its work. The company needs enough standardization to control costs and enough flexibility to fit very different warehouses.

The Cost of Making It Work

The financial record shows a company growing rapidly while still absorbing substantial costs.

Revenue increased from RMB807 million in 2023 to RMB1.36 billion in 2024 and RMB2.02 billion in 2025. First-half 2026 revenue reached RMB1.12 billion, up approximately 70% from a year earlier.

Gross margin improved from 16% in 2023 to 34.4% in the first half of 2026. The improvement suggests that growth has begun to translate into better economics at the project and product level. Financial overview

Profitability remains further away.

Reported net losses were approximately RMB1.01 billion in 2023, RMB1.26 billion in 2024 and RMB828 million in 2025, followed by RMB431 million in the first half of 2026. Those figures include financing-related accounting effects associated with investor redemption rights, so they should not be treated as a direct measure of operating cash consumption.

Even after adjustments, however, Hai remained lossmaking. Sales and marketing expenses reached approximately RMB587 million in 2025, compared with research and development spending of RMB385 million. The cost of finding customers and delivering a global business is a major part of the company’s financial burden. Latest filing analysis

Warehouse automation also creates a timing problem. Components, engineering and installation may require expenditure well before a project is completed and paid for. A growing order book provides visibility, but it also creates work that must be financed and delivered.

Hai’s filing indicates that it expects to remain in net loss for 2026. The company’s ability to turn growth into profitability will depend on project execution, overseas operating efficiency and control over the cost of each new installation. China Fund report

The IPO would supply capital for that next stage. It would also expose the company to a more persistent question: how much of its growth can become repeatable profit?

After the Robots Go Live

Hai Robotics has spent a decade expanding the usefulness of one relatively simple idea: retrieve the box, bring it to the operator and coordinate the journey through software.

Its development from early case-handling machines to climbing robots shows how much can be built around that premise. The robot changes first. The storage layout follows. Eventually, the entire warehouse operates differently.

That is what makes Hai an important company to understand. Its influence reaches beyond the number of machines it sells. It is helping customers reconsider how buildings, inventory and people should be arranged.

The limits of that influence are commercial. A warehouse operator must be convinced that the investment will pay back, that the system will survive changing order patterns and that its supplier will remain available years after installation.

Hai has established a growing international business and improved its gross margins. It now needs the organization around its machines to become as efficient as the warehouses it promises to automate.

The robots have learned to climb. The company’s next task is to make the economics follow.