Many humanoid robot companies tell the story of factories, and the logic is quite simple:
- Declining birth rates, an aging population, and a widening labor gap in factories lead to increased costs;
- Traditional automation solutions have high transformation costs and lack flexibility.
Therefore, as long as the issues of robot intelligence and cost are resolved, large factory clients can be secured, opening up a new market.This logic is theoretically flawless; there is demand and technology, and mutual matching can create a chemical reaction.However, business models do not rely on theory but on actual benefits.The flaw in this logic lies in the many prerequisites, such as assuming that robots possess high intelligence, that costs are low enough, and that there is demand. The reality is quite the opposite: factories do not have that much demand, costs are not attractive, and the notion of high intelligence is far from reality.First, the labor gap in factories is for high-end labor (skilled workers with strong learning abilities), not low-end labor.The downturn in real estate has released a large number of low-end laborers, numbering in the hundreds of thousands or even millions, whose skills do not meet factory needs.In purely manufacturing factories, there are many workers from the 60s and 70s doing the simplest jobs, working on assembly lines at extremely low costs. The current costs of robots, no matter how calculated, cannot compete with this group of people.Secondly, factories already have low gross margins, leaving even less profit space. A significant portion of factories derives profits from “exploiting” workers, such as extending working hours (three shifts, two shifts) and providing no benefits (no social security, no housing fund).A reality is that the dark factories operated by large companies are merely showrooms and lack universality. The vast majority of factories do not have the financial strength or willingness to automate production lines, let alone implement robots.Furthermore, factory environments are harsh, and the demands for robots are extremely strict, requiring both efficiency and stability. In addition, there must be low maintenance costs to reduce after-sales issues.Next, there are distinctions between private and state-owned factories; those that prefer to operate dark factories are mostly state-owned enterprises, as they need to maintain a facade. Private enterprises generally have little demand unless the owner is irrational. State-owned enterprises must go through a bidding process, which is lengthy and complicated.Finally, regarding payment models, factories generally do not engage in one-time purchases but rather make phased payments based on delivery, usage, and after-sales, making it difficult to collect payments and extending the cycle. Some factories even have additional clauses that deduct money if a certain phase goes wrong.It can be said that the intelligent transformation of factories is not voluntary at any step; it is all out of necessity.Many in the robotics field come from highly educated backgrounds and lack experience in dealing with traditional factory owners, making them prone to losses.Of course, stating this does not deny the factory scenario or disparage the customer base. Rather, it is essential to recognize the reality: factories are in this state, and robot companies must first consider how to integrate better into this environment.Here is a story from the mobile robotics industry as a reference.A logistics robot company had to accept cost prices to secure a project with a major manufacturer, leaving no profit margin. However, compared to having no profit, the greater fear was having no project.Operating at a loss can be PR’d as securing a key account, but without a project, there is truly nothing, making it difficult to explain to investors. Even if it means failure, it must be on a project basis.In fact, many companies initially adopt the mindset of “let’s just do it, a lower gross margin is fine” to at least secure the project. However, as they proceed, they realize that working with factories is not easy, slowly pouring resources into it, becoming more entrenched and deeper in the trap.The original intention of entrepreneurs may be to use technology to change society, filled with confidence to disrupt the industry, only to end up in a situation of “how did it come to this?”A project with dozens or hundreds of engineers stationed at the factory for delivery—can you handle that?Just the engineers’ salaries already exceed the project’s profit. They were excited to secure the order, only to find that once they started, there was no response. Subsequently, they reluctantly took on projects just to satisfy the organization.For humanoid robot companies, these are lessons learned from predecessors. Factories indeed represent a scenario with demand, but before entering factories, please first consider:
Is the demand that exists in factories truly in need of (our) solutions?
Is our solution the optimal one?
What is the purpose of undertaking this project?
Even if these questions are clarified, it only proves that current humanoid robot companies can undertake similar projects, but it is not advisable to do too many, as it may lead to “addiction”—falling into the ‘factory trap’, which is akin to drinking poison to quench thirst.As for the future, will robots disrupt the industry? They certainly will. But this cycle will be long, starting from ten years—how to survive that long without being eliminated is a very real and urgent question.
Previous Highlights 
Robot Communication Summary (2025.10)
Pengxing’s Old Stories
Ten Years of Seeking Shadows, Seeking Shadows for Ten Years