It is easy to make robots perform for show, yet getting them to work reliably inside real factories is an entirely different challenge. For 12 years, Hikrobot has focused on one core mission: deploying robots to handle actual production tasks on factory floors. These robots bear no resemblance to humanoid forms, yet they can observe surroundings and collaborate with one another just like human workers, delivering tangible value to manufacturing operations and people’s daily lives.
In 2026, at one side of the booth during the Intelligent Manufacturing Conference, Hikrobot showcased a wheeled embodied intelligent robot. Wheeled robots boast inherent advantages for factory scenarios, and this model has already taken charge of material handling tasks on Hikrobot’s internal production lines.
Hikrobot abides by a strict product principle: products will only be launched to the market once fully mature and validated for specific industrial scenarios. “We deliver tangible industrial products to customers, not empty technological visions,” said Robert Jia, CEO of Hikrobot.
Breakthroughs in algorithms such as reinforcement learning over the past two years have delivered unprecedented upgrades to robots’ motion control capabilities, enabling robots to execute complex physical movements with stronger environmental perception. Fueled by these technological leaps, a wave of startups focused on embodied intelligence has sprouted, sparking a nationwide frenzy over humanoid robots.
Still, staging robot demonstrations is far simpler than deploying them to work stably in real factories.
Originating as an internal incubation team under Hikvision back in 2014, Hikrobot spent nearly 12 years enabling large-scale robot deployment across industrial sites to generate genuine value for manufacturing.
The first half of this journey demanded patience and persistence: every product requires a 3–5 year R&D cycle, paired with repeated iteration and optimization of integrated hardware and software systems, plus validation across thousands of industrial sites with distinct workflows and on-site conditions. Hikrobot spent a minimum of five years to achieve its first large-scale commercial rollout.
By 2019, Hikrobot had shipped 1 million industrial cameras and over 10,000 autonomous mobile robots (AMRs) to market.
The second half of its journey brought explosive growth. China’s industrial upgrading wave unleashed massive market demand, and Hikrobot’s capacity to co-create robot-based problem-solving solutions with customers expanded rapidly across all sectors.
To date, cumulative shipments of Hikrobot machine vision products have exceeded 10 million units, while more than 180,000 AMRs have rolled off production lines. In China’s domestic market, one out of every two industrial cameras and one out of every three mobile robots is manufactured by Hikrobot.
Jia remains convinced this is merely the starting point. Speaking at the Intelligent Manufacturing Conference, he noted that manufacturing stands at a crossroads: emerging technological waves are reshaping supply capacity, while demand is shifting toward small-batch, high-variety, highly fragmented production.
Hikrobot has fully prepared for this shift. Its newly completed Tonglu production base is projected to hit full capacity in two years, and the company is scouting sites for additional manufacturing facilities.
In Jia’s vision, Hikrobot will evolve into a platform-based intelligent manufacturing enterprise serving manufacturing and logistics sectors, supplying all necessary intelligent hardware, software equipment and integrated systems.
Building Full-Stack Capabilities From Scratch
In 2014, an internal team led by Robert Jia was incubated within Hikvision, tasked with applying artificial intelligence and robotic technologies to industrial fields.
Circa 2014, two pivotal industry shifts unfolded. First, the fading demographic dividend drove rapid growth in China’s industrial automation, generating massive market demand. Second, the combination of convolutional neural network (CNN) algorithms, data and computing power unlocked revolutionary breakthroughs in AI, creating a window for Chinese enterprises to leapfrog global competitors.
Jia recognized that industrial intelligent upgrading is the only path to sustainable growth for Made-in-China, with AI set to become the core driving force of robotics. Manufacturing and logistics represent the most viable scenarios for rapid robotic deployment and value delivery.
From a technical standpoint, Hikvision boasted profound accumulated expertise in hardware, embedded development, ISP image processing and pattern recognition vision. At that time, mainstream products from overseas leading manufacturers still relied on outdated industrial pattern recognition algorithms, while Hikvision had already deployed cutting-edge CNN models for image recognition in security and commercial scenarios.
This technical edge led the team to believe it could penetrate the market via top-down technological innovation, similar to many internet and tech firms of the era. Yet the chasm between pure technology and genuine market demand became the first major hurdle the startup team needed to overcome.
In 2015, Jia led his team to develop three industrial cameras packed with innovative new features with full confidence.
One notable innovation was introducing color enhancement to industrial cameras — a function widely used in photography and security surveillance to produce human-friendly visuals. However, the team quickly uncovered a critical flaw during market rollout: most industrial vision systems feed data to algorithms, not human operators, eliminating the need for color rendering.
Unlike security and commercial applications, industrial scenarios prioritize stability far above cost. An industrial camera may cost merely a few thousand RMB within a production line worth hundreds of thousands, yet a single faulty camera can halt the entire piece of equipment.
“Customers will only be willing to replace existing equipment if new products deliver substantial tangible value,” Jia explained.
As a new market entrant competing against established players with decades of experience in vision recognition, what unique value could Hikrobot deliver?
Jia’s team landed on a clear answer: build everything from scratch.
Machine vision encompasses a complex ecosystem of hardware and software including industrial cameras and algorithms. Most new entrants opt to purchase off-the-shelf modules and focus solely on algorithm design. Hikrobot, however, resolved to independently develop nearly all machine vision components, from core algorithms to hardware and software systems.
For example, GigE Vision communication interface modules for industrial cameras demand ultra-stable data transmission. While many manufacturers purchase ready-made modules to cut development time, Hikrobot invested extensive time refining its in-house version, repeatedly debugging cross-protocol compatibility and universal adaptability.
On the hardware front, industrial cameras feature ultra-compact form factors, and the team spent years optimizing power consumption and heat dissipation within minimal physical dimensions.
On the algorithm front, Hikrobot pioneered AI algorithm-powered industrial barcode readers, triggering a generational leap in industrial code reading performance across the industry.
“Purchasing third-party modules accelerates product integration, yet it prevents deep reconstruction, optimization and system-wide iteration,” Jia said. “Without full control over individual modules, you cannot break free from existing technical frameworks. Plenty of 85-point products populate the market, but crafting a 95-point product poses immense challenges.”
Only products hitting that 95-point performance threshold deliver transformative value to customers.
This full-stack, ground-up development capability enables Hikrobot to optimize every modular component during product R&D, laying the foundation for its competitive edge across mobile robots and articulated robotic arms in subsequent years.
Co-Creation With Customers, Solving Real-World On-Site Pain Points
Robert Jia delivers structured, vivid speeches that balance rational analysis with illustrative metaphors — a reflection of his career trajectory. He was Hikvision’s first algorithm engineer, and later took charge of the group’s supply chain management.
During over a year in supply chain roles, Jia visited numerous lighthouse factories nationwide and oversaw the construction of Hikvision’s manufacturing base in Tonglu, Zhejiang. This hands-on experience granted him deep insight into manufacturers’ genuine demands. For instance, the most intractable pain point within many factory supply chains lies not in production itself, but intra-factory logistics.
Warehouse environments feature complex overlaps of personnel and goods, serving as critical links connecting upstream and downstream production. They form the weakest link in the manufacturing value chain, while also presenting one of the earliest viable scenarios for full intelligent transformation. For this reason, Jia’s team developed AMRs as a parallel product line alongside machine vision: vision systems act as the factory’s intelligent “eyes,” while mobile robots serve as its intelligent “feet.”
At that time, the market already offered various material handling equipment such as automated guided vehicles (AGVs), yet these devices suffered two universal limitations. First, constrained by outdated algorithms and hardware, they could only travel along fixed pre-defined paths. Second, equipment manufacturers lacked deep understanding of industrial scenarios; factory logistics involves complex on-site conditions requiring intimate knowledge of cross-industry production workflows.
Optimizing existing hardware could not generate incremental value for factories — the core priority was understanding scenarios and solving practical problems, a gap AMR systems were designed to fill.
In 2015, Hikrobot’s intra-logistics solution was validated and tested at Hikvision’s Tonglu manufacturing base, where the first batch of underride AMRs was developed. In January 2016, Hikrobot deployed its first large-scale AMR project at the Tonglu plant, rolling out 800 underride robots in a single installation. Inside its own factory, the AMR system endured rigorous real-world production pressure and iterative refinement. Deployment to automotive plants and fresh food warehouses followed later.
In 2017, a supermarket retail client faced steeply rising labor costs, low sorting efficiency and high error rates within its fresh food distribution center, creating urgent demand for intelligent transformation. The client opened its warehouse for joint trials despite Hikrobot’s limited prior experience in fresh food scenarios. Through continuous trial and error, the two parties deployed 40 AMRs and seven sorting workstations across a 4,000-square-meter fresh food warehouse. The workflow shifted from “workers traveling to goods” to “goods delivered to workers,” lifting sorting efficiency from 120 pieces per person per hour to 210 pieces.
This customer co-creation model defined Hikrobot’s early development, with the express delivery industry serving as a typical case study. Back in 2017, almost no domestic vision brands operated in logistics; parcel sorting, code reading and weighing relied entirely on manual PDA scanners. Logistics firms sought to develop domestically tailored DWS (Dimension-Weigh-Scan) systems and partnered with Hikrobot for joint R&D.
The primary technical hurdle for DWS in logistics lies in deformed shipping labels stuck on irregular parcels, often covered with transparent adhesive tape that impairs code reading. Given the extreme complexity of real-world sorting lines and minimal global precedent, overseas leading vendors largely avoided this market, targeting only high-budget clients with clean, standardized scenarios. Domestic logistics companies turned to local intelligent manufacturing firms like Hikrobot for viable solutions.
To accumulate data and test systems, the logistics partner reserved dedicated sorting lines exclusively for Hikrobot’s development team. Algorithm engineers worked onsite from sweltering summer to frigid winter, spending months completing initial development. Post-launch, the team spent years ongoing optimization before the solution saw widespread industry adoption in 2019.
After 2019, Hikrobot onboarded countless new clients across emerging industries including automotive, lithium battery, photovoltaic, semiconductor and medical devices. Executives at these manufacturers readily embraced robotics, and capacity expansion accelerated demand for automated equipment. New factories were designed with dedicated space for large-scale robot deployment from the ground up.
Your message must be between 20-3,000 characters!
