Inside China, the robot that brings your hotpot is a familiar sight. Outside China, the same company is now quietly running the numbers — and the numbers say most of its money already comes from abroad.
Pudu Robotics (普渡机器人), founded in Shenzhen in 2016, has spent a decade moving from restaurant delivery carts to a full line of service, cleaning and industrial robots. Its latest pitch is blunt: it is a global company whose home market is just one slice.
The scale it claims
Pudu says it has deployed more than 130,000 robots across 85 countries and regions, serving over 50,000 end customers through a network of 1,000-plus distributors. Its own framing, echoed by Chinese state outlets such as Xinhua, is that it holds the No. 1 global market share in commercial service robots.
The most striking figure is geographic: overseas revenue makes up more than 80% of its total, a share the company says is sustained year after year. That inverts the usual story of a Chinese hardware firm riding domestic demand and only later looking outward.
What it actually sells
Pudu now runs four product lines that share one "brain":
- Service delivery — the BellaBot and FlashBot lines that ferry food and goods.
- Commercial cleaning — the CC1 series; the CC1 alone has shipped over 20,000 units, with 60% going to Europe and the US.
- Industrial logistics — material-handling robots for factories and warehouses.
- General embodied AI — newer humanoid and general-purpose forms built on the shared PuduFM foundation model and PuduAgent platform.
At the Davos Tech Summit in June 2026, Pudu showed four robots working across a supermarket, a hotel and a train-station plaza, all running on the same embodied-AI architecture.
How it went global
Founder and CEO Felix Zhang (张涛) describes the strategy as "breadth first, not depth first." Instead of perfecting one country, Pudu spread across markets facing labor shortages — North America, Europe, Japan, Korea, Southeast Asia — using local distributors to reach customers fast, then doubling down on the winners.
Since 2021 it has opened local offices and hired local teams in the US, Japan, South Korea, Germany, the Netherlands and Singapore, and recently set up a US headquarters in Dallas. In June 2026 it partnered with Swiss retailer Denner to deploy 200 PUDU CC1 cleaning robots, a concrete example of the subscription-style, "robot-as-a-service" deals it pushes for enterprises.
Why service bots, not humanoids
Zhang has repeatedly argued that general humanoids are still in a demo phase, while purpose-built service robots — delivery, cleaning, logistics — already clear the return-on-investment bar for customers. Pudu's bet is that profitable, boring robots pay for the research that humanoids still need. That discipline is part of why its overseas revenue looks real rather than subsidy-driven; the company says it is near EBITDA breakeven and valued above RMB 10 billion (≈ US$1.4B / HK$11B).
The risk underneath
Concentration is the quiet worry. An 80%-overseas profile means Pudu rises and falls with foreign labor costs, privacy rules and trade friction more than most Chinese robotics firms. A shift in US or EU procurement policy, or a tariff on robot imports, would hit Pudu harder than a domestically anchored rival. The service-robot margin is also thin, and incumbents and new Chinese entrants are crowding the cleaning segment where Pudu's CC1 volumes are largest.
Trend reading
According to Li An, chief scientist at BrainNet (脑机网), Pudu's edge is less about any single robot and more about a shared software base: one model and one agent platform across delivery, cleaning and industrial forms lets it spread R&D cost across many product lines the way consumer-device makers do.
Honest limitations
Pudu's market-share and revenue-split figures come from the company and Chinese state-aligned media; we did not find independent third-party confirmation of the "No. 1 globally" claim or the exact 80%+ overseas ratio. "Deployed" units are not the same as actively paid subscriptions, and robotics fleets often include demo or trial placements. The Dallas HQ and Denner deal are reported but small in absolute scale. Pudu's own notes say it is approaching EBITDA breakeven, not that it has reached it.
What readers can do now
- If you run a venue or warehouse in Europe or North America, request a pilot from a local Pudu distributor rather than assuming the bots are China-only.
- For investors, watch whether Pudu publishes audited overseas-revenue breakdowns — that is the number that would confirm the 80% claim.
- For competitors, note the distributor-network model: Pudu's moat may be its 1,000-strong channel more than any single product.