--- title: "China's Pudu Ships 120,000 Service Robots and Closes In on Profit" date: 2026-10-08 category: Humanoids & Robotics site: NeuroAI canonical: https://neuroai.site/a/na-robot-pudu-service-robots language: en --- # China's Pudu Ships 120,000 Service Robots and Closes In on Profit > Shenzhen's Pudu has put more than 120,000 service and cleaning robots into 80-plus countries and says its EBITDA is near break-even — a rare profile in a sector hooked on prototypes. A hotel guest in Frankfurt watches a small wheeled robot glide down the corridor, balancing a tray of amenities toward room 412. In a Dallas warehouse, another unit ferries parts between shelves without a single human in sight. The company behind both just became one of the few robotics firms in the world that can talk about *profit* instead of just prototypes. Pudu Robotics (普渡机器人), founded in Shenzhen in 2016, is not a household name outside China — but its machines are. In April 2026 the company closed a new funding round of nearly **RMB 1 billion (about $140 million / HK$1.09 billion)**, pushing its valuation above **RMB 10 billion (roughly $1.4 billion / HK$10.9 billion)**, according to reporting by [Pandaily](https://pandaily.com/pudu-robotics-secures-nearly-rmb-1-b-in-funding-and-expands-global-footprint-with-new-u-s-headquarters) and [RobotsBeat](https://robotsbeat.com/pudu-robotics-raises-150-million-series-valuation-1-5-billion/) that mirrors the company's own disclosure. Cumulative funding now exceeds **RMB 2 billion (around $281 million / HK$2.18 billion)**. ## Why 'boring' cleaning became the moneymaker The headline number is not the funding. It is the trajectory. Pudu says its revenue grew **more than 100% year-on-year in 2025**, and that its *EBITDA* is close to turning positive — which would make it one of the first commercial service-robot companies to approach profitability at scale. The engine behind that growth is unglamorous: floor scrubbing. Pudu entered commercial cleaning in late 2021 and, within four years, that segment grew to account for **more than 70% of total revenue**. Its CC1 cleaning robot alone has shipped **over 20,000 units**, roughly 60% of them into the compliance-heavy markets of Europe and North America. A robot that sweeps, mops, vacuums and scrubs in one pass fits the one thing businesses actually pay for: a repetitive, predictable task done cheaply and without absenteeism. ## The global footprint Pudu's cumulative shipments have passed **120,000 units**, reaching **more than 80 countries and regions**, over 1,000 cities, and **40,000-plus customers**, the company states. Its machines have landed in the supply chains of *Carrefour*, *Walmart*, *EDEKA* and *Tesco*. That global reach is not accidental. Overseas sales have made up **more than 80% of revenue for consecutive years** — a deliberate hedge against China's price-war domestic market. In April 2026 the company opened a U.S. headquarters in Dallas, Texas, and says it has already deployed roughly 15,000 robots across the Americas with 285% year-on-year revenue growth there. According to *Frost & Sullivan*, Pudu held a **23% global market share**, ranking first worldwide in commercial service robots (a 2023 report the company cites). The deployment variety is wider than the big-box names suggest. Pudu's FlashBot Pro delivery robot has been installed at *Atour (亚朵)* hotels in numbers exceeding 1,000 units, and the same model was placed at the Everest Base Camp at 5,200 meters — a stunt, but one that demonstrates environmental adaptability few service robots can claim. ## From delivery cart to 'one brain, many bodies' Pudu started with restaurant delivery carts — the kind that weave between tables carrying plates. It has since branched into four product lines: service delivery, commercial cleaning, industrial logistics, and general *embodied AI (具身智能)*. The strategic bet is architectural: a *"one brain, multiple embodiments" (一脑多形)* design that reuses navigation, scheduling, manipulation and interaction software across very different physical forms, from a low delivery bot to a humanoid. That reuse shows up in practical fixes. Pudu's semi-humanoid *Lightning Box Arm (闪电匣Arm)*, launched in March 2025, carries a manipulator arm whose job is simply to press elevator buttons — solving the *elevator-control (梯控)* barrier that blocks cross-floor robots in overseas buildings where retrofitting is forbidden. It is a small, dull engineering answer to a real deployment wall, not a leap of autonomy. The company says it holds **over 1,800 patents** and runs R&D centers in Shenzhen, Chengdu and Hong Kong. It has also begun shipping humanoids — the D9, unveiled in April 2025 with a self-developed "bionic cerebellum" (*PUDU PULSE*) — though humanoids remain a small, forward-looking slice of the business rather than the profit center. In August 2026, research firm Counterpoint separately ranked Pudu first worldwide in commercial cleaning-robot shipments and revenue. ## The honest read on the hype Pudu's story is often told as a China-versus-the-world robotics race. The more useful lesson is narrower: in service robotics, the clearest path to real money today is not the humanoid everyone films, but the *floor cleaner* nobody films. Structured, high-frequency, low-judgment tasks convert into purchase orders; flashy manipulation still converts into demos. That said, Pudu is not yet declaring a net profit, and its humanoid ambitions pit it against far larger, better-funded rivals. The April 2026 round was explicitly aimed at embodied-AI R&D and manufacturing scale — a sign the company knows the next contest is software, not hardware. ## Honest limitations Most of the financial figures — 120,000 units shipped, 80-plus countries, 40,000 customers, 23% market share, 2025 revenue growth above 100%, cleaning's 70% revenue share, and the near-break-even EBITDA — come from the company's own disclosures and from market-research claims by *Frost & Sullivan*, as relayed by Pandaily, RobotsBeat and Chinese financial media. They are not independently audited in this article. The valuation and funding amounts are consistent across multiple outlets but remain company-announced. "Near break-even EBITDA" is a directional statement, not a published audited figure. Overseas-revenue share and U.S. deployment counts are company-stated and not separately verified. The competitive ranking by *Frost & Sullivan* reflects that firm's methodology, not a regulator's view, and the August 2026 Counterpoint claim is a vendor-cited market report. ## What readers can do now - If you run a facility, ask vendors for the *total cost per cleaned square meter over a full year*, not the sticker price — Pudu's scale suggests the economics, not the demo, are what separate winners. - Watch the cleaning-and-logistics segment, not just humanoids, when judging which robotics firms will survive: recurring, structured tasks pay bills. - Treat any "valuation" headline in this sector as a funding signal, not a profit signal — confirm audited revenue and margin before reading it as success. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-robot-pudu-service-robots Free to quote with attribution and a link to the original.