A cleaning robot scrubs a supermarket floor in Texas. Another carries a tray down a hotel corridor in Florida. Both were designed and built thousands of miles away, in Shenzhen. The question facing the company that made them is no longer whether it can ship machines abroad. It is whether it can keep them running when something breaks at 2 a.m.
On April 23, 2026, Pudu Robotics (普渡机器人) opened a new United States headquarters in Dallas, Texas. The announcement, carried by the company and by PR Newswire, marked a shift in how a Chinese commercial service robot (商用服务机器人) maker sells overseas: less exporting, more local infrastructure.
What the Dallas office actually is
The new headquarters is not a factory and not a flagship store. According to the company's press release, it sits in Richardson's Sherman Tech Center and combines office space, a product showroom and on-site warehousing. Pudu also converted its Santa Clara office into a logistics support outpost and set up a dual-warehouse system on both US coasts.
That combination matters more than the ribbon-cutting. A robot company selling into hotels, hospitals and warehouses needs a place to hold spare parts, train technicians and demonstrate a machine that will be judged on uptime rather than novelty. Dallas, the company argues, gives it a central base for both North and South America.
The figures the company published
Pudu entered the US market in 2018. In the April release it reported:
- nearly 15,000 robots deployed across the Americas;
- regional revenue growth of 285% year over year;
- growth of 63.6% year over year in distributor partnerships;
- more than 120,000 units shipped worldwide, in over 80 countries and regions.
The same release cited a Frost & Sullivan study from 2023 that puts Pudu's share of commercial service robotics at 23%, and listed Walmart, Accenture, NASA, Norwegian Cruise Line and Honeywell among its customers. These are company-reported figures. The revenue growth is a percentage, not a disclosed dollar amount, so the underlying base is unknown.
In its April release, Pudu quoted Americas general manager Raymond Pan saying the company was building for the long term in the Americas with a localized approach.
What "localization" means when the building fights back
The phrase going global (出海) is common in Chinese tech coverage. The less glamorous version is localization (本土化), and in service robotics it involves unglamorous obstacles.
Chinese financial press describing Pudu's overseas work reported that many foreign office buildings do not open their lift systems to third-party software. Rather than negotiate an API that may never come, the company built robots that press physical buttons, the way a person would. Where a site cannot be rewired for water and drainage, the cleaning machines carry their own supply and recovery tanks. These are mechanical workarounds, not AI breakthroughs, and they are often what decides whether a deployment survives.
The product matrix sold in the Americas spans four lines: BellaBot and BellaBot Pro for delivery; the CC1, MT1 and BG1 series for cleaning; T-series industrial carriers with payloads from 150 kg to 600 kg; and the D5 series for what the company calls general embodied intelligence (具身智能).
Europe is the other half of the story
The Dallas announcement was not the only overseas move. A Pudu release dated September 14, 2026, said the company was deepening its European presence after IFA Berlin, where the PUDU D7 made its first European appearance. That release reported more than 130,000 robots shipped globally and operations in more than 85 countries, with business in over 20 European countries.
It also named specific deployments: Denner, part of Switzerland's Migros Group, deployed 200 CC1 cleaning robots across stores; Pudu robots are used at MediaMarkt, Albert Heijn, Carrefour Poland and EDEKA in Germany. The company said it plans a new European headquarters in Germany. Both releases use 'more than' counts — above 120,000 in April and above 130,000 in September — so the direction is consistent, though the exact increment is unknown.
The "one brain, multiple bodies" pitch
Underlying the expansion is a technical claim the company calls it one brain, multiple forms (一脑多形): a shared model, described as PuduFM, and an agent platform, PuduAgent, driving robots of different shapes. Specialized machines handle repetitive tasks, semi-humanoid units work in complex industrial and commercial sites, and humanoids are aimed at more general collaboration. It is a way to reuse software across a product catalogue rather than train each body from scratch.
Honest limitations
Every number in this article comes from Pudu's own press releases or from media repeating them. The 285% revenue growth has no disclosed base, no currency and no independent audit; the 23% market share comes from a 2023 Frost & Sullivan report cited by the company, and I have not read the underlying study. Customer lists are company claims.
Pudu is a private company, so there is no public filing to check profitability, cash flow or how much of that growth came from acquisitions. The word "leader" appears in the company's own releases and should be read as marketing, not as a neutral assessment.
The localization details — lift buttons, self-contained water tanks — come from Chinese financial reporting based on company accounts, not from independent site visits. None of the releases say what a robot costs to buy or to keep running for five years. Service quality varies by country; a headquarters and two warehouses do not prove that a technician will arrive the same day in Brazil or Poland.
What readers can do now
- If you run a hotel, hospital or store: ask for a paid pilot on your own floor, with your own lifts and drainage constraints, not a showroom demo. Uptime and recovery time matter more than the robot's vocabulary.
- If you are comparing suppliers: request the failure behaviour in writing — what the robot does when it cannot dock, who is notified, and how fast a human arrives.
- If you follow the sector: watch service infrastructure — local warehouses, technician counts and distributor depth — rather than shipment totals. Those are what turn a Chinese export into a lasting local business.
