What Changed After I Walked Through Unitree’s Factory
Last Wednesday afternoon, I stood inside Unitree’s factory in Yuhang, Hangzhou, watching four-legged robots get assembled and tested. Before the visit, I assumed robotics entrepreneurship was mostly a contest of capital and supply chains. Afterward, I kept wondering whether Western indie teams could enter through a specific use case at a lower cost.
What Kind of Opportunity Is This?
Unitree is not selling a vision of the future; it is selling relatively standardized, more affordable four-legged robots that people are already using for inspections, performances, and R&D. Instead of building hardware from scratch, a small team can validate demand with an off-the-shelf machine. I used to get stuck building the product first and finding customers later. Eventually, I realized that leasing or buying a machine and selling a service around it is often a steadier starting point.
My friend Ajie demonstrated an inspection project inside a logistics park in Foshan in March 2025. He did not build a robotics team. He bought an existing machine, landed a small night-shift park-patrol project, then provided a human service based on recordings and anomaly alerts. That project showed me the opportunity is not only in selling robots. It is also in solving a problem around the robot that a customer will pay for.
What It Costs to Replicate This Path Today
Buying or leasing one machine costs RMB 20,000–100,000, with another 1–4 weeks set aside for testing the site, interfaces, and staff training. The technical barrier is understanding what the machine can do and communicating effectively with suppliers and engineers; writing low-level control code is not required. The first step is straightforward: open two established suppliers’ websites or contact their sales teams, choose a four-legged robot, click “Book a Demo” or “Get a Quote,” and ask about delivery, after-sales support, and secondary-development fees.
This tool is not for everyone, and there is no urgency to test it now. If you have no orders or industry contacts, or your customers do not yet have a clear use case, stocking equipment can trap cash. Starting with manual inspections and documenting the problems before deciding what to automate is safer.
Choosing a Path for Your Current Stage
If I were just starting, I would first spend RMB 500–2,000 to rent a machine, collect three real tasks, and avoid rushing to incorporate a company or develop hardware. With one or two customers already in hand, I would choose a repetitive, dangerous, or night-shift scenario and quote a machine fee plus a service fee, then run one paid pilot. When scaling, I would put after-sales support, spare parts, operators, and delivery across regions into the cost model before discussing bulk purchases. Another option is to build a business around training, inspection operations, or scenario integration rather than manufacturing the robots yourself.