On August 3, DJI’s patent infringement lawsuit against Insta360 over its Luna series went to trial at the Shenzhen Intermediate People’s Court. The two sides presented evidence, cross-examined each other’s submissions, and delivered arguments. The hearing concluded without a verdict, and it remains unclear when a ruling will be issued.
A separate dispute is also being heard over DJI’s claim that six patents held by Insta360 were service inventions created by former DJI employees and therefore belong to DJI.
The dispute began in late March, when media reports said DJI had filed lawsuits with the Shenzhen court seeking ownership of six Insta360 patents. Two days later, Insta360 founder Liu Jingkang wrote on WeChat that he planned to turn the competitive tactics his company had encountered over the previous year into a novel. The protagonist was widely understood to allude to DJI founder Frank Wang.
Over the following months, the patent war moved from the Shenzhen court to the US District Court for the Eastern District of Texas, then back to China. During that period, Insta360’s Luna Ultra gimbal camera and DJI’s Pocket 4P went on sale in the same month, with neither product gaining a decisive edge in market reception.
Yet the headquarters of the two companies are only about a 30-minute commute apart. From DJI Sky City to Insta360’s offices at Qianhai Life Insurance Financial Center, a taxi runs east along Liuxian Avenue before turning south onto Nanping Expressway, covering just over ten kilometers.
The courtroom, retail shelves, and social media are all battlefields. But the combatants share the same supply chain and draw from the same pool of engineers.
That kind of geographic proximity is normal in Shenzhen’s hardware industry.
In June, AI² Robotics, which in February became Shenzhen’s first embodied intelligence unicorn valued above RMB 10 billion (USD 1.5 billion), raised nearly RMB 5 billion (USD 744.5 million) more, doubling its valuation to more than RMB 20 billion (USD 3.0 billion). On the same day, X Square Robot also announced that its valuation had surpassed RMB 20 billion.
AI² Robotics described itself as the first embodied intelligence company in the Greater Bay Area to officially exceed a valuation of RMB 20 billion. X Square Robot called itself the first and only embodied intelligence company in the area with a valuation above that threshold.
The two are only five kilometers apart.
Products, lawsuits, and financing all point to the same underlying structure: within a roughly 50-kilometer radius in Shenzhen, supply chains, talent pools, and capital networks have been compressed into an extraordinarily dense space.
Shenzhen is arguably the only city in the world where an idea can be turned into a product and then produced in large volumes on such a short cycle. It has a remarkably complete supply chain, extensive manufacturing capacity, rapid prototyping capabilities, and extremely fast product iteration. In many other places, an idea might remain a PowerPoint presentation. In Shenzhen, two weeks later, it may already be a prototype.
In 2025, Shenzhen’s total industrial output and value-added industrial output among enterprises above designated size remained among the highest of any Chinese city for the fourth consecutive year. The city had 1,333 national-level specialized and sophisticated “little giant” enterprises, the most nationwide. Its robotics industry generated RMB 242.6 billion (USD 36.1 billion) in output, up 20.56% year-on-year, while Shenzhen led the country in industry scale, number of companies, listed companies, and financing events.
In November, the Asia-Pacific Economic Cooperation (APEC) forum will once again put Shenzhen in the global spotlight. The 33rd APEC Economic Leaders’ Meeting will take place in Shenzhen from November 18 to 19, making this the third time China has hosted the meeting, after Shanghai in 2001 and Beijing in 2014. International political leaders will discuss global trade rules in Xiangmihu.
Across Shenzhen, different districts play distinct but highly coordinated roles in the hardware economy: frontier R&D, brand strategy, mass manufacturing, global expansion, and experimentation with future technologies. Capital rotates rapidly between sectors, while companies take on markedly different characteristics depending on their stage of development.
What to build, how to build it, and how to scale it: different parts of Shenzhen answer different industrial questions.
Behind those questions are different groups of people: job seekers assessing opportunities, investors selecting projects, and supply chain partners looking for collaborators.
Which corporate culture, neighborhood, and stage of capital correspond to which way of working?
Six locations, one ecosystem
Open a map of Shenzhen and the locations of its hardware companies are anything but random.
Six major areas occupy six different ecological niches. Some produce technology, some produce brands, and some produce manufacturing capacity. What matters is not just the individual companies, but their positions relative to one another and how those positions determine the way they obtain resources.
Nanshan
The 12-kilometer Liuxian Avenue connects Xili Lake International Science and Education City, the headquarters base at Liuxiandong, and the Tanglang industrial community.
Along this corridor sit clusters of hardware companies including DJI, Anker, EcoFlow, Shokz, Lumi United Technology, AI² Robotics, and LimX Dynamics. The corridor is also an important part of Shenzhen Robot Valley.
As a hub for Shenzhen’s robotics industry, Robot Valley has attracted more than 200 companies across the robotics supply chain and 32 specialized and sophisticated national-level “little giant” enterprises. It is also home to 11 leading universities and research institutions, including Southern University of Science and Technology, Harbin Institute of Technology, Shenzhen, and Tsinghua Shenzhen International Graduate School.
There is a simple way to classify companies along this corridor: whether their roots trace back to DJI.
Anker and LimX Dynamics do not.
Anker moved the bulk of its R&D operations in October 2025 into a new building at China Resources Snow Breweries’ science and technology complex in Xingdong, Bao’an. Its previous offices were in Building B of Jianxing Technology Building. LimX sits across the road from Anker’s former location, just a few hundred meters away in a straight line. It completed a pre-IPO funding round in July at a valuation of around RMB 15 billion (USD 2.2 billion).
Both companies wanted to remain close to the talent pool around Xili University Town without being pulled into Liuxiandong’s competition for expensive office space.
The center of gravity, however, is Liuxiandong. And that is DJI territory.
In 2006, Hong Kong University of Science and Technology student Frank Wang rented a warehouse of less than 20 square meters in Shenzhen to work on flight control systems. His professor, Li Zexiang, became an angel investor.
DJI later shifted from selling flight control components to designing complete products in-house. In 2013, its Phantom series took drones beyond the model-aircraft community and into the mass consumer market, eventually capturing roughly 70% of the global consumer drone market.
Today, DJI Sky City is a Liuxiandong landmark. It cost RMB 1.6 billion (USD 238.3 million) to build. DJI has filed more than 18,000 patent applications, more than 60% of them for invention patents.
But DJI’s most important contribution to Nanshan may not be its own revenue. It is the talent system that continually spills outward.
DJI experienced frequent employee departures in 2025, with several core sales staff joining competitors. Sources said the company had not placed some of those employees under noncompete agreements, allowing them to bring DJI’s methods and strategies directly to rivals.
Insta360, seeking to quickly close technical gaps in core drone technologies such as flight control, video transmission systems, and obstacle avoidance, began hiring aggressively at high salaries. Its R&D headcount jumped 59% in 2025, including several core R&D employees who had left DJI.
Li Zexiang once compared technology entrepreneurship to walking at night.
“You’re scared when you’re alone,” he said. “But if there are two or three of you, or three or five, and someone gives you a flashlight and a stick to fend off dogs, you’re no longer afraid.”
That road is now crowded with people from the DJI ecosystem.
Across from DJI Sky City, inside Chuangzhi Cloud City, EcoFlow founder Bruce Wang left DJI’s battery division in 2017 and went on to build the world’s largest brand offering portable energy storage solutions. The founding team of consumer 3D printing leader Bambu Lab came from DJI’s consumer drone division. WizEvo Tech, Tars, AgileX Robotics, HEQ UAV Tech, and several other companies also carry talent from DJI.
InnoX, founded by Li himself, is also based in Chuangzhi Cloud City. XbotPark, the incubator he helped build, produced LiberLive, a smart guitar startup whose founder previously worked on navigation algorithms at DJI.
Working at DJI is, in essence, buying into a long-term belief in technology.
The company puts technology first and is willing to give engineers working on foundational R&D substantial room to experiment and fail. But complaints about unclear promotion paths and intense overtime frequently appear on platforms such as Maimai and Xiaohongshu.
That creates a paradox: DJI continually produces founders who go on to build startups worth tens of billions of RMB, yet the same culture of uncompromising R&D that makes DJI so productive also helps push some of its talent toward entrepreneurship.
For venture capitalists hunting for the next DJI-linked startup, the vicinity of Liuxiandong offers a good chance of finding clues.
About five kilometers east of Shahe West Road is Nanshan Zhiyuan’s Chongwen Park in Tanglang, an office complex designed to attract major projects in integrated circuits, artificial intelligence, the internet, universities, and research institutes. It covers about 179,000 square meters.
The industrial ecosystem here can be compressed into a single elevator.
In the first building, the eighth floor houses Lumi United Technology, a smart home product maker that grew out of the Xiaomi ecosystem. Its R&D team numbers 460 people, or 48.9% of its workforce.
The 19th floor houses AI² Robotics, an embodied intelligence unicorn valued at RMB 20 billion.
Eric Guo, also known as Guo Yandong, holds a doctorate from Purdue University and founded AI² Robotics in April 2023. Guo has experienced the full cycle of the smart device boom, having previously worked as a researcher on Microsoft’s US AI team and later served as chief scientist at Xpeng and Oppo.
At the start of his entrepreneurial journey, he focused on something that had yet to become mainstream: an end-to-end vision-language-action model that maps robot perception directly to action without an intermediate layer.
In June 2024, AI² Robotics open-sourced GOVLA0.0, or RoboMamba, which it described as the first VLA model open-sourced by a startup globally. Its self-built production line entered operation in September 2025, and the company signed a three-year, 1,000-unit-scale equipment order with HKC, the world’s third largest display panel manufacturer. The agreement ranks among the larger commercial contracts for productive robots in the industry.
In 2026, the company further upgraded its technology with the NeuroVLA brain-inspired model and open-sourced AlphaBrain, an ecosystem platform for embodied intelligence.
At the 2026 World Robot Conference, AI² Robotics moved away from making stage demonstrations its main showcase. Instead, it deployed AlphaBot2 in the event venue’s official merchandise store to take real customer orders.
The same model can make coffee and ice cream and perform industrial tasks such as picking up printed circuit boards.
The emphasis was on practical deployment: robots should not just take the stage. They should be put to work.
If GOVLA can prove commercially viable across a wide range of use cases, the stock options held by AI² Robotics’ early employees could become considerably more valuable. But if the technology fails to gain market acceptance, the company’s heavy upfront R&D investment could prove costly.
Its team combines academic and engineering backgrounds. Its R&D work follows experimental and evaluation standards comparable to those used at top AI conferences, while a self-built computing cluster containing thousands of accelerator cards supports model training. Its campus recruitment spans algorithms, motion control, multimodal systems, and hardware engineering.
About one kilometer east along Shahe West Road, at 4221 Xilihu Road, is Shokz’s R&D center.
Its registered headquarters and manufacturing center are in Shancheng Industrial Zone in Shiyan, Bao’an, but its R&D team is concentrated around Xilihu Road in Nanshan. Such arrangements were once the classic layout for Shenzhen hardware companies.
Honghualing Base is changing that assumption. Located northwest of the Liuxian Avenue corridor, the base officially opened on February 24, 2026. Its first group of tenants included 14 leading companies and specialized players, among them RoboSense, Dobot, EngineAI, and SMiT.
It is a flagship example of Shenzhen’s verticalized factory model: 16-story industrial buildings, spiral ramps that allow trucks to drive directly to the ninth floor, and a floor area ratio as high as 6.71.
EngineAI’s Honghualing production line began operating in May, producing one T800 humanoid robot every 15 minutes. RoboSense shipped nearly one million LiDAR (light detection and ranging) units in 2025 and 330,000 in the first quarter of 2026 alone, up more than 200% year-on-year.
Estimates show that once all 14 initial tenants reach full production, their combined output over five years will exceed RMB 183 billion (USD 27.2 billion) in economic value.
R&D can remain in Liuxiandong while mass production is compressed into a radius of just one kilometer or so. Nanshan’s industrial chain is gradually closing the loop.
Follow Shahe West Road south toward the coast and you reach Yuehai Subdistrict. Across roughly 14 square kilometers, 103 listed companies sit in this subdistrict, or more than seven per square kilometer.
Tencent’s Binhai Building and the Baidu International Building are neighbors. Xiaomi’s Shenzhen international headquarters and ByteDance’s building sit across the street from one another.
Hardware companies are similarly compressed. Inside the same Shenzhen Bay Innovation and Technology Center, EngineAI occupies the 26th floor and Astribot the 68th. The two humanoid-robot unicorns share an elevator. Bambu Lab’s registered headquarters is also nearby, in the United Headquarters Building.
Liuxiandong is only about a 20-minute drive away.
Zoom out and a pattern becomes obvious.
In Nanshan, the radius of talent movement is extremely small. Changing jobs within the same district does not require changing apartments. It may not even require changing cafeterias.
A three-kilometer radius around Liuxiandong is enough to serve as prime territory for recruiters seeking relevant talent.
Nanshan is the heart of Shenzhen’s hardware industry and one of China’s most important centers for consumer hardware and embodied-intelligence innovation.
Compared with other parts of Shenzhen, its industrial culture is more aggressive. Traditionally, the central questions companies here discuss are not yield rates, capacity, or delivery schedules.
They are technical pathways, the limits of iteration, and industry standards.
Bao’an
Bao’an’s hardware landscape follows a clear contour line.
Close to Qianhai, Bao’an’s central district is filled with brand headquarters. Move north, and the terrain rises as factories grow denser. By the time you reach Shiyan, you are in the heartland of Shenzhen’s traditional manufacturing economy.
In one tower of Qianhai Life Insurance Financial Center on Xingye Road in Xin’an Subdistrict sits Insta360’s headquarters.
Founder Liu Jingkang moved the company from the Yangtze River Delta to Shenzhen in 2015. The main reason was a bottleneck in the mass production of optical modules for panoramic cameras.
Shenzhen’s comprehensive supply chain provided a complete solution, from drawings to production.
Insta360 listed on the Star Market in June 2025. Full-year revenue reached RMB 9.7 billion (USD 1.4 billion), up 74.76% year-on-year. According to IDC, it held 66% of the global panoramic-camera market and 57% of the thumb camera market. Net profit attributable to shareholders of the parent reached RMB 929 million (USD 138.3 million).
The key choice separating Insta360 from traditional hardware companies has been its refusal to rely exclusively on heavy offline distribution. Instead, it has pursued content-led globalization.
Through offline brand exposure at CES, viral distribution on overseas social media, and an AI editing engine trained on data from millions of users, it turned panoramic cameras from a nonexistent category into one in which it became the global market leader.
Its longstanding confrontation with DJI is, in a sense, an unavoidable part of its evolution from a niche player into an all-out platform competitor.
Compared with DJI’s serious and regimented corporate culture, Insta360 has taken a more relaxed approach as the challenger. That has helped fuel its growth and its appeal to talent.
According to China Entrepreneur, Insta360 has a highly flat organizational structure. Employees are given substantial room to make mistakes in their daily work, while internal transfers can span very different roles and are relatively easy.
The company also regularly holds internal competitions, supports employees incubating new brands, and even encourages workplace relationships.
Still, Insta360’s RMB 9.7 billion in revenue in 2025 came with a problem: sales increased, but profits did not. Third-quarter revenue rose 92.64% year-on-year. For the full year, revenue grew while net profit attributable to the parent fell 6.62%.
By the first quarter of 2026, revenue had reached RMB 2.5 billion (USD 372.3 million), up 83.11% year-on-year, but net profit attributable to the parent had fallen 52.02% to RMB 84.6 million (USD 12.6 million).
Yuan Yue, head of Insta360’s China business, has said publicly that the company ranks employee development first, customer value second, and shareholder returns third. “It’s not that we aren’t worried. There is pressure,” he said. “But we can’t let financial figures distort our actions.”
Insta360’s employees are 28 years old on average. Execution is extremely fast, the business spans the globe, and the career value of having the brand on a resume continues to rise as the company expands.
The tradeoff is that its training system and salary structure remain underdeveloped amid rapid growth, while work is fast-paced and long-term pressure is high.
Whether its internal management systems can keep up with its expanding scale is a central question for employees weighing a job there. It is also one of the capital market’s lingering concerns.
Qianhai Life Insurance Financial Center sits within the sphere of influence of the Qianhai Shenzhen-Hong Kong Modern Service Industry Cooperation Zone.
In the industry’s early days, robotics companies often chose to locate in lower-rent Xili. But one industry source who has visited the Qianhai Authority told 36Kr that many companies are now voluntarily moving into more expensive parts of the area.
They are attracted by government services, targeted industry subsidies, and early access to industrial application scenarios.
Qianhai is developing a dedicated plan for the robotics sector, though detailed rules have not yet been released publicly.
In July, a China center for APEC cooperation on cross-border data flows was officially announced for Qianhai.
That elevated cross-border data flows from a local Qianhai pilot into a multilateral issue involving Asia-Pacific economies. The center is a flagship outcome of China’s 2026 APEC year and was jointly established by the Cyberspace Administration of China and the Shenzhen municipal government. It is responsible for policy research, regulatory alignment, and industry services.
Leveraging its proximity to Hong Kong, Qianhai has already tested pioneering initiatives, including Shenzhen-Hong Kong cross-border data verification. Its related DDTP, or distributed data transmission protocol, verification model has also been submitted as a Chinese proposal for discussion under APEC’s digital economy mechanisms and has received responses from multiple parties.
Follow Bao’an’s industrial contour north and you reach supply chain giants such as Luxshare Precision and Avary Holding. Shokz’s full production operation is tucked among 26 factory buildings in Shancheng Industrial Zone in Shiyan.
A consensus has formed within Shenzhen’s hardware community: for frontier technology partnerships, go to Nanshan. To check production capacity, yield rates, and delivery schedules, go straight to Shiyan.
Those in Shiyan do not usually participate in top-level strategic decision-making. But they directly determine whether products can be delivered reliably.
Xin’an, Fuyong, and other subdistricts also host substantial manufacturing capacity that turns Nanshan’s R&D output into physical products.
Compared with Nanshan’s role as an innovation source, Bao’an combines brand headquarters with a production hinterland.
Insta360 and Woan represent mature consumer brands that have already validated their technologies. Their central challenge is scaling and global operations. Shiyan, meanwhile, handles mass production.
Qianhai is an economic functional zone with its own statutory body, the Qianhai Authority, an agency dispatched by the Shenzhen Municipal Government that is responsible for policy, investment promotion, and development.
Basic administrative functions such as household registration, subdistrict administration, civil affairs, and policing remain under Nanshan and Bao’an districts. It is a typical model in which an administrative district and an economic zone are managed separately.
Futian
Futian’s hardware industry has two poles. At the top is Honor’s headquarters. At the foundation is Huaqiangbei’s components market.
They are only eight kilometers apart, yet represent three generations of industrial development.
Honor’s three office locations in Futian together form a complete map of its organizational structure.
At the northernmost point, Shenzhen New Generation Industrial Park in Shangmeilin houses its core R&D organization. This was Honor’s first campus after it separated from Huawei in November 2020. The transfer of assets and move-in of its full workforce took just two months, supported throughout by a dedicated working group organized by the Futian district government. Parts of buildings one through four house R&D teams. Honor held 11,466 valid invention patents in 2024, accounting for 25.7% of Futian’s total.
Travel five kilometers south and Honor’s global headquarters comes into sight at Shum Yip UpperHills in Xiangmihu.
It officially opened in August 2021. The company’s registered headquarters and global marketing and sales operations are concentrated here.
The logic behind the division is straightforward: keep R&D close to talent and supply chains, while placing the global headquarters in the city center to establish the brand at the top of the hierarchy.
To the east, the Shenzhen-Hong Kong Cooperation Zone for Technology and Innovation in Hetao handles frontier exploration and has become a key base for Honor’s shift toward an AI-device ecosystem.
Honor set up an R&D laboratory there in 2023. It later became one of the core pillars of the Honor Alpha Plan, under which the company plans to invest more than USD 10 billion over five years across products including smartphones, tablets, and robots.
At the 2026 MWC Barcelona, Honor unveiled its first humanoid robot and the Robot Phone.
At Beijing’s robot half-marathon in April, Honor Robotics took first place with its D1 robot, while its A1 received the “Best Gait Award.”
On August 22, D1 broke the human record in the 100-meter event at the second World Humanoid Robot Games.
A1, which is aimed at the consumer market, is scheduled for small-scale production in 2026 and is expected to be deployed first in Honor stores to attract customers and provide shopping assistance.
Honor’s core team inherited Huawei’s talent development system and compensation framework. Strong state ownership provides patience for long-term strategy, while its IPO counseling process continued in 2026, with the fundraising size still undetermined.
But the disadvantages of a large organization are equally apparent. A 200-person robotics business has limited influence inside an organization employing tens of thousands. Decision-making chains are long, and competition for internal resources is routine.
Geographically, the frontier R&D base sits in Hetao, the decision-making center remains in Xiangmihu, and execution teams are rooted in Shangmeilin.
The efficiency of commuting between those three points directly affects how long a new project takes to move from approval to implementation. It also determines how efficiently the voices of frontline employees travel upward.
Eight kilometers east of Xiangmihu lies the capillary network of Shenzhen’s hardware ecosystem: Huaqiangbei.
At a counter just one meter wide, buyers can find every component needed for a circuit board. Solution providers, surface-mount factories, and mold makers can complete the entire production relay within a handful of buildings.
Shenzhen’s reputation for rapidly turning ideas into prototypes depends on Huaqiangbei’s grassroots ecosystem.
Honor represents the top of Shenzhen’s hardware pyramid, controlling high-level strategy and global brand influence.
Huaqiangbei represents the foundation, providing every entrepreneur with components, rapid prototyping, and lightweight, low-cost hardware solutions.
As Shenzhen’s central business district, Futian is not built for large-scale contract manufacturing facilities.
One pole determines where the industry is going. The other determines how quickly an idea can take physical form.
Longgang
Huawei is the dominant gravitational force in Longgang’s hardware landscape.
Centered on Tian’an Cloud Park, the entire Banxuegang Science and Technology City has grown around Huawei.
Construction on Huawei’s Bantian campus began in 1997. Nearly three decades later, Huawei behaves like a planet, pulling the whole Bantian area into its orbit.
The resumes of local entrepreneurs often follow a remarkably similar pattern: leave the Huawei system, then start a business by drawing on its mature upstream and downstream supply chain.
It is Longgang’s most distinctive entrepreneurial path.
Tian’an Cloud Park now hosts nearly 1,000 high-quality companies and more than 50,000 technology workers. Nearby are Huawei’s consumer device operations and its Bantian campus.
A community centered on one-person companies, or OPCs, launched in 2026 and operates across roughly 3,000 square meters of incubation space. It has already attracted nearly 100 individual entrepreneurs and teams.
Across Banxuegang, 614 national high-tech enterprises and 287 specialized and sophisticated companies have established operations. The main sectors are security, communications, electronics, and intelligent hardware.
Compared with Nanshan’s aggressive innovation culture, Longgang companies tend to be more pragmatic. Survive first. Talk about frontier innovation later.
Absen in LED displays, Sekorm in electronic component distribution, and Chinasoft International in software services are among the companies embedded to varying degrees in Huawei’s upstream and downstream ecosystem.
Even when consulting firms such as Accenture move into the area, one of their core objectives is to capture the benefits created by Huawei’s industrial concentration.
But Bantian is now developing a second growth curve beyond the Huawei ecosystem.
On July 28, 2025, the world’s first robot “6S store” opened at Galaxy World in Bantian.
The “6S” model expands on the traditional automotive framework of sales, spare parts, service, and customer feedback by adding equipment leasing and personalized customization.
Nearly 70 robotics brands moved in at launch. Demonstration units from Leju Robot, Unitree Robotics, LimX Dynamics, EngineAI, and other companies are displayed side by side, combining demonstrations and commercial testing in one location.
With the Galaxy World store at its center, Longgang has established itself as a district for embodied intelligence within roughly one square kilometer.
Longgang now hosts more than 12,700 companies across the AI and robotics supply chain, or roughly one-fifth of all companies in Shenzhen’s robotics cluster.
Longhua
Longhua’s structure is relatively simple: Dalang handles original design and original equipment manufacturing for finished devices, Guanlan houses large intelligent-manufacturing plants, and Yousong contains both Foxconn’s enormous industrial complex and office buildings occupied by emerging consumer-hardware brands.
The baseline topics of daily conversation among Longhua companies are production schedules and monthly shipment volumes.
More specifically, Dalang is a global production center for mini PCs.
Major manufacturers of finished devices, including AZW Technology, CYX Industrial, and GMK Technology, are concentrated here. Their products reach consumers worldwide through cross-border e-commerce.
Factories operate at or near full capacity for much of the year. Although these companies do not control how much end consumers pay, they directly influence global mini PC price ranges.
Guanlan, meanwhile, hosts large intelligent-manufacturing companies such as Sunlord Electronics and Envicool. It is one of Shenzhen’s most comprehensive manufacturing districts in terms of supporting industries and overall production capacity.
Two different industrial models coexist in Yousong.
Foxconn’s Longhua campus is an industrial city unto itself, cementing Longhua’s contract manufacturing identity. In office buildings in the same area, companies such as Sharge have pursued differentiated niches in consumer hardware.
Longhua is also an important pillar of Shenzhen’s 3D printing industry.
Creality is headquartered in Minzhi Subdistrict. Elegoo is based in Dahe Industrial Zone in Guanhu Subdistrict.
Founded in 2015, Elegoo sells directly to overseas consumers through its own website and Amazon. Its products reach more than 150 countries, with overseas markets generating more than 90% of revenue.
Add Bambu Lab in Nanshan and Anycubic in Longgang, and four Shenzhen companies together control roughly 90% of the global consumer 3D-printing market.
The headquarters of the two farthest apart are still less than an hour’s drive from one another.
Guangming and Pingshan
At Shenzhen’s northern edge lies Guangming. To the east is Pingshan.
Together, they perform the long-term groundwork for the city’s future.
Guangming is accumulating foundational frontier science and technology. Pingshan is refining advanced automated-manufacturing capabilities.
These two districts represent the next decade of Shenzhen’s hardware industry.
Guangming Science City is a core base for Shenzhen’s future industries. It hosts major scientific facilities focused on synthetic biology and brain analysis, as well as more than 140 synthetic-biology companies with a combined valuation approaching RMB 42 billion (USD 6.3 billion).
In consumer hardware, Guangming’s key areas are brain-computer interfaces and hardware for senior care and wellness.
Another major advantage is its university cluster.
Phase one of Sun Yat-sen University’s Shenzhen campus is now in operation, while Shenzhen University of Advanced Technology has been formally approved.
The goal is to create a complete talent loop in which people study locally, start companies locally, then sell their products globally.
In Guangdong’s 2026 college entrance examination admissions, Harbin Institute of Technology, Shenzhen, had a minimum admission score of 662 for applicants to its physics track. Southern University of Science and Technology’s AI program group had a cutoff of 656. Newly established Shenzhen University of Advanced Technology reached 629.
Shenzhen universities’ ability to attract high-scoring students continues to reinforce the city’s role as a base for frontier science and education.
Pingshan, by contrast, is a showcase for high-end intelligent manufacturing.
Honor’s Intelligent Manufacturing Industrial Park is located on Lanzhu West Road and covers some 135,000 square meters. Here, a smartphone rolls off the line every 28.5 seconds on average. Automated systems perform 85% of processes, and more than 60% of the equipment was developed by Honor itself.
Pingshan is not simply pursuing maximum production capacity. Its role is to test the upper limits of automated manufacturing. Products requiring high precision and highly automated mass production come to Pingshan to prove that they can be built.
The district is simultaneously developing four major industries: automotive, pharmaceuticals, semiconductors, and intelligent technologies.
Industrial output among enterprises above designated size has exceeded RMB 600 billion (USD 89.3 billion).
At their current stage, Guangming and Pingshan are windows into longer-term industries whose commercialization will generally take time.
The six areas now form a complete relay chain for Shenzhen hardware.
Huaqiangbei turns ideas into prototypes quickly. Shiyan provides stable yields and reliable delivery. Longhua supplies global shipment volumes. Pingshan pushes the limits of intelligent manufacturing.
Technology originates in Nanshan. Brand headquarters and top-level decision-making sit in Bao’an and Futian.
Long-term industrial possibilities are being stored in Guangming.
From a design drawing to a global mass market launch, the entire supporting chain fits within a 50-kilometer radius.
Three levels of funding, three different fates
As discussed earlier, geography determines how companies approach resources. Capital is the most visible of those resources.
Nanshan is closer to technology, making it easier for companies there to win high valuations on the strength of a technical path that has not yet been fully validated.
Bao’an and Longhua are closer to mass production, so companies there are more likely to gain investor recognition from the certainty of shipment volumes.
When investors sit in a Xili coffee shop screening projects, they are not merely comparing teams and products. They are also considering whether a company’s location gives it the conditions needed to support its current capital market proposition.
A company’s prospects, and the scope for others to work with it, depend on more than where its office is located. Its position in the capital cycle matters just as much.
Listed companies need to solve the problem of finding a second growth curve. Unicorns need to scale and commercialize. High-potential startups need to build differentiated barriers in niche markets. Early seed-stage companies need to prove that their underlying technical path works at all.
No stage is inherently better than another. Every financing position contains both opportunity and risk.
Unicorn companies
AI² Robotics and LimX Dynamics are both coveted by investors, but capital evaluates the two companies in fundamentally different ways.
In theory, AI² Robotics’ GOVLA model can bridge industrial and household scenarios. But there is an obvious weakness behind its high valuation: the model depends heavily on data from real-world environments.
The company’s current revenue and data accumulation remain concentrated in industrial applications. Although it has expanded into public service and retail scenarios, it has not yet entered the consumer home market.
Following its Series B financing in February 2026, AI² Robotics surpassed a valuation of RMB 10 billion. After raising nearly RMB 5 billion in another round in late June, its valuation exceeded RMB 20 billion. The company has completed joint stock restructuring and begun preparations for a Hong Kong IPO.
Investors are willing to value it above RMB 20 billion not because of current performance, but because they are betting on whether it can survive the long, cash-intensive period before its investment in data begins to work.
LimX Dynamics is taking a different path. It built its reputation around motion control and is now moving toward a full-stack architecture that integrates the robot’s “large brain” and “small brain.”
Its applications have also expanded beyond its early focus on industrial inspection into research and education, commercial services, and all-terrain inspection. Across its full product portfolio, it has accumulated orders for several thousand units, more than half from overseas customers.
Its financing has moved just as quickly. It raised USD 200 million in a Series B round in February 2026, followed by nearly USD 200 million in a pre-IPO round in July, giving it a post-money valuation of about RMB 15 billion.
The market is willing to pay that price because LimX Dynamics has grown outward from the single capability of motion control into a full-stack platform.
Neither path is inherently superior. But they imply different rules for survival.
AI² Robotics must continually prove that its technology can generalize across scenarios.
LimX Dynamics must show that the full-stack capabilities it has built from a single technical strength remain reliable and deliverable as it expands into more applications.
Bambu Lab, Creality, Anycubic, and Elegoo, Shenzhen’s “four little dragons” of 3D printing, together account for about 90% of the global consumer 3D printing market. Yet three of the four remain unlisted.
Bambu Lab is the largest of them. Using high-speed FDM, or fused deposition modeling, printing and its multicolor AMS, or automatic material system, Bambu Lab turned 3D printers from niche toys for enthusiasts into more general-purpose household tools.
It remains private and has no public record of filing a prospectus. Its last publicly confirmed financing was a Series C round in April 2024. Its valuation in the 2023 Series B round was about USD 2 billion.
Another unlisted company, Honor, is far larger than a conventional unicorn.
Reports have put Honor’s pre-IPO valuation at RMB 200 billion (USD 29.8 billion). The 2026 Hurun Global Unicorn Index valued it at RMB 180 billion (USD 26.8 billion), making it one of the most closely watched unlisted assets in Shenzhen.
Part of that valuation reflects the steady earnings generated by its established smartphone business. Another part reflects the longer-term potential of humanoid robots, the Robot Phone, and AI devices.
Honor has already assembled a 200-person robotics team. But educating consumers about humanoid hardware will be expensive, and the technology is still not mature enough to support mass consumer adoption.
How much of Honor’s RMB 200 billion valuation should ultimately be attributed to the potential of these new businesses will depend on how capital markets judge the timeline for consumer humanoid robots.
The company’s listing process itself has no clear schedule. Honor filed for A-share IPO counseling in June 2025. A counseling inspection initially expected in March 2026 has passed its expected date, and no firm listing timetable has been announced.
Standing at the edge of the unicorn tier is Lumi United Technology, which has filed for a listing and already has one foot over the threshold.
The company formally filed an application with the Hong Kong Stock Exchange on March 29, 2026, but has not yet listed. Its post-Series D valuation was USD 850 million.
Over the past several years, its gross margin rose nine percentage points across three years, while the Aqara brand has become partly independent of the Xiaomi Home ecosystem.
Before filing, the company fully repurchased its Series D preferred shares in three transactions, spending USD 51.3 million and allowing early institutional investors to exit in advance.
Going public
Going public does not reduce the number of problems a company must solve. It simply changes the question from one of survival to another focused on turning an existing advantage into the next phase of growth.
UBTech Robotics illustrates this. It captured an early advantage through a conventional architecture that separated the robot’s “large brain” from its “small brain.” But in 2025, its share of global humanoid-robot installations was only around 5–6%, as newer players such as Unitree and AgiBot continued to eclipse it in shipments and public attention.
At the end of June 2026, the company launched the U1 biomimetic companion robot. Preorders exceeded 10,000 units as it attempted to open a new consumer growth market.
But rebuilding an older technical architecture requires substantial engineering work, while simultaneous R&D for industrial and household applications continues to consume cash.
The market still recognizes its first-mover advantage. But investor caution is growing.
The battle in 3D printing, meanwhile, is a classic business story of new technology disrupting traditional scale.
Creality has chosen to defend through scale. It listed on the Hong Kong Stock Exchange on May 29, becoming Hong Kong’s first publicly traded consumer 3D-printing company.
Its shares opened more than 80% higher on the first day before retreating from their intraday peak. In its first-half 2026 results, the company swung from profit to loss.
Its Ender series has shipped millions of units. It later introduced the K1 series to confront growing competitive pressure from Bambu Lab.
Unlike Creality, which must now answer to the public market, Bambu Lab remains privately held and is judged by a different set of expectations.
The key question is not shipment volume, but whether 3D printing can move beyond a niche hobby and become a mainstream necessity. That is the shift the market is waiting for.
Anker and Insta360 are also undergoing transitions, but their problems center on channels and branding, respectively.
Both companies place overseas markets and global branding at the center of their strategies. But the capital market is less concerned about their ability to go global than about whether their existing models can become durable operating capabilities.
For years, Anker’s story was often reduced to a single line: reduce dependence on Amazon and build a global multichannel brand.
In July, it completed a dual listing in Shenzhen and Hong Kong, bringing that transformation narrative under scrutiny in two capital markets.
Yet its financial statements suggest that years of transition have reduced its dependence only modestly. The channel resilience investors want to see has not yet fully materialized in verifiable operating results.
Insta360 followed a different route.
It began online, using overseas social media content and its self-developed AI editing tools, and in recent years has accelerated investment in physical retail.
It now has more than 300 branded stores worldwide and entered more than 500 Apple authorized reseller stores in 2026. Online and offline revenue are now close to evenly split.
At its post-listing peak, Insta360’s market capitalization exceeded RMB 100 billion (USD 14.9 billion). By the end of last month, it had fallen below RMB 50 billion (USD 7.4 billion), down roughly 70% from the peak. Its first-half 2026 net profit plunged by more than 90% year-on-year.
Investors are now watching to see whether the attention that fueled Insta360’s early growth can translate into lasting brand value.
Startups with potential
Sharge and Zuowei, a nursing robot company also based in Longhua, are representative examples.
Neither has publicly disclosed a valuation, and each has raised funding totaling hundreds of millions of RMB. But their situations are different. One continues to iterate and raise capital under the shadow of industry giants. The other has yet to establish a replicable business model.
For companies at this stage, any delay in commercialization can immediately make the next round of financing difficult because they generally lack access to the very largest pools of capital.
The capital market is showing an extreme Matthew effect. A small number of companies considered to have greater certainty are being fiercely pursued, while a large number of midtier businesses remain stuck at commercialization bottlenecks.
Startup teams still bear the cost of choosing the wrong technical path. At the same time, the overall level of industry financing continues to rise.
In China’s 2025 unicorn rankings, Shenzhen was among the leading cities nationwide in the number of newly added unicorns.
Finding your coordinates in a compressed city
Capital markets have never benefited every player equally.
In the first half of 2026, embodied intelligence companies in the Guangdong-Hong Kong-Macao Greater Bay Area raised about RMB 4 billion (USD 595.6 million) in total.
But from the second half of 2026 through 2027, three barriers could start to eliminate midtier companies: choosing the wrong technical route, failing to commercialize applications as expected, and coming under pressure from weak cash flow.
Companies that survive over the long term may need to meet three conditions:
- A clear and commercially viable technical path.
- Applications validated in real markets.
- Sufficient cash reserves.
At this moment in Shenzhen’s hardware industry, listed companies are competing on operating efficiency and resilience against the risks of globalization.
Unicorns are competing on the speed of real-world deployment.
High-growth companies are competing on differentiated positioning.
Startups are competing on their judgment of foundational technologies.
Within the same Shenzhen industrial belt, companies at different coordinates have already embarked on dramatically different trajectories.
But the fascinating thing about Shenzhen is that those trajectories remain geographically concentrated.
Technology obsessives may be best suited to the atmosphere of the Liuxiandong corridor in Nanshan, where people talk about academic papers, computing clusters, and industry standards. But they must also accept unclear promotion paths and intense overtime.
Professionals focused on global expansion can find content-driven brands such as Insta360 in Qianhai, but they need to weigh whether personal-development systems can keep pace with rapid corporate growth.
Those seeking greater stability and lower volatility may find more room for error among supply chain companies in Longgang’s Huawei ecosystem.
And people willing to bet on longer-term industries can find opportunities in brain-computer interfaces in Guangming and intelligent manufacturing in Pingshan. But those opportunities may require time and patience.
Inside this compressed, 50-kilometer industrial belt, changing jobs can mean changing sectors, company types, and even life trajectories without changing cities.
Service providers, entrepreneurs, investors, and employees can all choose a position that matches their appetite for risk.
And it is amid this intense competition at such close quarters that some people ultimately find a long-term place for themselves.
KrASIA features translated and adapted content that was originally published by 36Kr. This article was written by Fang Tingyu for 36Kr.
Note: RMB figures are converted to USD at rates of RMB 6.71 = USD 1 based on estimates as of September 23, 2026, unless otherwise stated. USD conversions are approximate and, where appropriate, rounded for ease of reference. They may not fully match prevailing exchange rates.