The upper limit of robots lies in Beijing, and their survival depends on Guangdong.

If you review the primary market for robots in the past year, you'll witness a highly interesting "cognitive fold."

In the first half, the spotlight was indeed on Haidian in Beijing and Zhangjiang in Shanghai. Back then, investors shuttled between cafes around universities, discussing the pros and cons of the Transformer architecture and the boundaries of generalization ability. At that stage, everyone firmly believed in one thing: Software defines everything, and the model is the barrier. As long as the "brain" is strong enough, the "body" is just an executor.

However, in the second half of 2024, with the implementation of the State Council's "Opinions on Deeply Implementing the 'Artificial Intelligence +' Initiative" in August, a touch of cold rationality suddenly mixed into this wave of enthusiasm.

When the demo turns into a PPT, and the PPT needs to become a mass - production plan, the industry's wind vane has subtly shifted. Investors and entrepreneurs have alarmingly discovered that to embrace the national - level trillion - dollar benefits and move from the "prototype" in the laboratory to the commercial "blockbuster product," there lies an insurmountable physical chasm:

How can those algorithms that perform perfectly in the simulator be downloaded onto a body that must cost less than $20,000 and can withstand the harsh real world full of oil and dust?

When the problem changes from "being smarter" to "being more durable" and "being cheaper," the focus of China's robotics industry inevitably begins to shift southward.

Looking across the Yangtze River and beyond the Nanling Mountains, our sight finally focuses on the approximately 50 - kilometer arc - shaped industrial belt from Nanshan in Shenzhen to Shunde in Foshan. There are no excessive metaphysical discussions about "silicon - based life" here, only extreme calculations regarding BOM (Bill of Materials) cost, NPI (New Product Introduction) cycle, and yield rate.

If Beijing and Shanghai define the "intellectual ceiling" of Chinese robots, then this 50 - kilometer area in the Pearl River Delta is propping up the "floor" for the implementation of China's robotics industry through its strong engineering capabilities.

Algorithms are not a panacea; the physical world has its "quirks."

Why is the strategic position of this "50 - kilometer area" being re - evaluated today?

Because the biggest misunderstanding in the humanoid robot industry currently is the excessive superstition of the "end - to - end" large model, believing that as long as there is enough data, robots can develop general abilities like ChatGPT. However, front - line experts from the Institute of Automation of the Chinese Academy of Sciences and the Tsinghua Institute for Interdisciplinary Information Sciences have pointed out an embarrassing reality for pure - algorithm companies - the "illusion of simulation data."

In the computer simulation environment, the friction of a cup is constant, the lighting is uniform, and the ground is flat. But in the real world, the cup may be greasy, the light may refract, and the ground may have just been mopped by a cleaner. A grasping algorithm that runs perfectly in the simulator may not even be able to hold a wet glass firmly once it is downloaded onto a physical robot.

This is the famous "Sim2Real (from simulation to reality)" gap. Crossing this gap cannot rely solely on writing code; it must rely on the "body's" feedback.

This is the irreplaceable underlying logic of the Pearl River Delta. It can not only manufacture motors but also provide a vast amount of real "physical feedback data."

On this 50 - kilometer corridor, more than 8,500 robot - related enterprises are clustered. This is not just a cold number but a huge neural network.

The high - precision sensors of Inovance Technology and Tongchuan Precision can capture micron - level force feedback, telling the large model "whether it's holding an egg or a stone." The visual radars of OPPO and Leishen Intelligence train the robot's "eyes" under extremely complex factory lighting conditions. More importantly, 70% - 80% of the supply chain here overlaps with that of new - energy vehicles. The battery thermal management technology of BYD and the wiring harness technology of Luxshare Precision, which have been verified millions of times in vehicles, are being directly applied to robots.

Algorithms need a body to sense the "quirks" of the physical world, and the most sensitive, strongest, and cheapest body can only be manufactured most efficiently in the Pearl River Delta. This is no longer a simple division of labor between "southern factories and northern stores" but a deep - seated evolution of "decoupling and re - constructing software and hardware."

Shenzhen: "Dual - wheel Drive" of Policy and Market

After the industrial chain solves the problem of "whether it can be done," the next challenge is efficiency. In the hardware industry, the speed of NPI (New Product Introduction) often determines survival.

Currently, the form of humanoid robots is not yet finalized, and the standards are not unified. It is in a period of high - frequency trial and error. At this time, the core competitiveness of enterprises is not "holding back a big move" but "changing quickly." The core value of Shenzhen lies in providing a globally rare super - verification environment of "policy support + market acceleration."

On the one hand, policies such as the "22 Measures for Prioritizing Manufacturing in Guangdong Province" clearly identify embodied intelligence as a strategic fulcrum for future industries, giving the industrial chain a sense of security. On the other hand, Shenzhen's unique "money - making" culture makes the local supply chain instinctively sensitive to new technologies.

An entrepreneur who moved from Beijing to the south once sighed to me: "In the north, I need to spend half an hour explaining to suppliers what embodied intelligence and multimodality are. In Shenzhen, the supplier boss interrupts me directly and asks, 'Can this product be mass - produced next year? If so, I'll make samples for you for free, betting on your future.'"

This kind of conversation occurs in Shenzhen on an hourly basis.

There is a very fine - grained division of the supply chain here. Components such as frameless torque motors and coreless motors, which were once "choke - points" dominated by foreign companies, are being rapidly localized in thousands of small, medium, and micro - enterprises, CNC machining centers, and PCB rapid - prototyping factories in Shenzhen. The reason why Ubtech can secure the world's largest single - order (over 800 million yuan for the Walker series) and DJI can maintain or even lead the global iteration speed is essentially because they have "built" their R & D laboratories on the supply chain.

This engineering efficiency advantage based on geographical location, combined with policy dividends, makes Shenzhen the first choice for new product implementation. Here, verifying an idea doesn't require a long process; you just need to turn left out of the door and have a chat with a boss.

Dongguan: The Best Soil for Replicating the "Tesla Model"

If Shenzhen solves the problem of "fast verification," then moving north along the Guangzhou - Shenzhen Expressway, Dongguan is responsible for solving a more brutal problem: "difficulty in mass production."

Elon Musk once claimed that he would bring the price of Optimus down to $20,000. This figure is like the Sword of Damocles hanging over the heads of all Chinese enterprises, forcing them to think about two questions: How to increase production capacity? How to halve the cost?

We must face the data honestly: Currently, the planned annual production capacity of leading domestic enterprises such as Ubtech is still in the level of 100,000 units, which is still orders of magnitude behind BYD's single - factory production capacity of millions. In terms of manufacturing scale and standardization, we still have a long way to go to reach the ideal "Tesla model."

However, Dongguan is the best, and perhaps the only, soil for narrowing this gap.

There is a large amount of production capacity that has been tempered by both the "Apple supply chain" and the "automobile supply chain." There are not only 5,000 mold factories but also a methodology for "turning precision products into low - cost items."

Enterprises such as Yinbaoshanxin and Changying Precision are trying to transfer the "stamping + die - casting" process of automobile manufacturing to robots. This "cross - industry reuse" ability is crucial. As an expert from CDH Investments said, "70% - 80% of the supply chain for the manufacturing of embodied intelligence entities can be shared with new - energy vehicles." Dongguan's large - scale automobile parts industrial cluster is ready to provide low - cost structural components for humanoid robots at any time.

In this sense, Dongguan is not just a factory; it is a "compressor" for BOM costs.

Precision components such as planetary ball screws and six - axis force sensors, which were once out of reach, are undergoing technological breakthroughs in Dongguan's precision factories. Whoever can secure Dongguan's production capacity will survive the future price war. This is understood by new car - making forces and is also beginning to be understood by unicorn robot - making companies.

Foshan: The "Value Anchor" for Defoaming

The end - goal of the industrial chain is application. For B - end customers, robots are not only carriers of technology but also means of production.

We cannot avoid the current bubble in the primary market: A large number of projects are concentrated in the seed and angel rounds, with pre - valued but delayed revenues; the PPTs look beautiful, but the commercial verification is insufficient. The only answer to digesting these bubbles is: Enter the factory, do the work, and get the payment.

Foshan is the "value anchor" for defoaming.

As the heartland of China's manufacturing industry, Foshan has world - class factories such as Midea, Haier, and FAW - Volkswagen. The bosses here don't listen to stories or watch cool demo videos. They only look at two data in the Excel spreadsheet: MTBF (Mean Time Between Failures) and ROI (Return on Investment).

In Geely's 5G smart factory, Ubtech's Walker S2 is already working in collaboration. In some unmanned warehousing centers, the wheeled robots of Shidou Intelligence are trying to replace human labor. The screening criteria here are extremely practical, even brutal.

A factory director in Foshan once said bluntly to me: "Don't talk to me about high - tech. If this machine 'acts crazy' on the production line or breaks down once a day, I'll kick it out of the workshop in two days."

The environment here is full of dust, oil, vibration, and strong electromagnetic interference, which is a nightmare for delicate electronic components but a compulsory course for mature industrial products. Currently, several robots such as those from Ubtech and KUKA are being piloted in Foshan factories. This cruel feedback of "scenario - driven R & D" is the inevitable path for China's embodied intelligence to move from PPT - based financing to a commercial closed - loop.

An investment manager said bluntly: "If a company can run through the process of orders, delivery, and payment in an industrial scenario like Foshan within one or two years, its valuation can be supported by fundamentals." Foshan's factories are using the strictest "test papers" to select truly viable enterprises.

Finally

When we raise our sight and re - examine this industrial map, we will find that the so - called "Beijing/Shanghai vs. Pearl River Delta" is not a zero - sum game but a relay race.

This is no longer a simple geographical division of labor but a deep - seated collaborative evolution of China's hard - technology industry.

National - level policies have pointed the way. The top - notch research teams in Beijing and Shanghai have broken through the algorithm ceiling of the "brain" and "cerebellum," solving the problem of "thinking like a human." The 50 - kilometer area in the Pearl River Delta has picked up the relay baton, using its strong supply - chain resilience and manufacturing heritage to cast a strong, cheap, and mass - producible "body" for these algorithms.

The future of embodied intelligence must be a combination of the "strongest brain" and the "strongest body."

In this national - level technological competition, the gap in algorithms may be narrowed by open - source communities and time, but the manufacturing barriers and the thickness of the industrial chain require decades of accumulation to form.

When the industry shifts from "concept verification" to "large - scale application," those closer to factories and supply chains have a greater chance of survival. On this map, the 50 - kilometer area in the Pearl River Delta is the "crucial leap" for Chinese robots to become a trillion - dollar industry. It is not only the starting line but also the line that determines the outcome.

Note: This article is written based on the publicly available version and industry analysis. Some functional details may be adjusted with product iteration.

This article is from the WeChat official account "Technology Must Not Be Cold," written by Balang and edited by Fangshan. It is published by 36Kr with authorization.