Xiaoyubot bets on one brain, many robots

One AI brain, many robot bodies—Xiaoyubot bets the future of welding on a single core.

Xiaoyubot’s founder, Qiao Zhongliang, embodies what observers call the “Xiaomi-school” lineage of embodied intelligence entrepreneurs. A Beihang University alum (class of 2010) who spent 13 years at Xiaomi—ultimately as head of MIUI R&D for five years—Qiao left to found Xiaoyubot in February 2023. The company is described in reporting as a startup focused on “industrial general-purpose embodied intelligence,” with welding robots as its inaugural product line. The ownership structure isn’t publicly disclosed; public reporting labels Xiaoyubot as private, with no explicit notes of state backing at this stage.

The core concept is captured in the Chinese phrase yi nǎo duō xíng (一脑多形): “one brain, multiple forms.” In practice, a single multimodal embodied intelligence brain would drive different robot morphologies—ranging from a single-arm unit to wheeled dual-arm configurations, and even humanoid forms—without rebuilding the control stack for each form. Qiao argues that once cross-hardware adaptation is achieved, the software-to-hardware integration burden could shrink to roughly 10 percent when moving to a new morphology. That “hit product logic”—borrowed from his Xiaomi playbook—aims to lock in a scalable, platform-based automation core that can be amortized across forms and tasks.

The market logic aligns with a sector-wide bottleneck: welding remains in short supply across China, with Qiao citing a shortage in the welding occupation category well over 10 million workers. That gap creates a sizeable arena for automation, especially in high-mix, high-variation welding tasks common in automotive, consumer electronics, and appliance manufacturing. Xiaoyubot’s play is not a one-off gadget but a platform proposition: secure top-tier components and push volume to drive unit costs down, then reuse the same brain across multiple morphologies to address a broad set of welding scenarios.

A milestone reported by the founder is that smart Welding robots entered customer production lines in the fourth quarter of 2025—a meaningful step from concept to factory floor, and an implicit validation of the “one brain, many forms” thesis. The deployment speaks to a broader trend: Chinese robotics startups increasingly blend software-driven intelligence with modular hardware to accelerate time-to-value on the factory line.

From a policy and ecosystem perspective, Xiaoyubot sits in a space where private, tech-enabled automation startups proliferate alongside state-led industrial programs. The public record doesn’t disclose state backing, but the model—leveraging software cores to drivetrain hardware across forms—resonates with a national push to upgrade manufacturing through platformized automation and AI.

Two-pronged implications for global manufacturers and Chinese suppliers emerge. First, if Xiaoyubot’s model scales, it could compress the traditional per-task customization cost curve, pushing OEMs and contract manufacturers to embrace more generalized automation cores rather than bespoke, task-specific cells. Second, a successful platform approach could tilt supplier dynamics toward software-led integration and modular hardware—from sensors and drives to control architectures—favoring players who can deliver end-to-end, scalable solutions rather than one-off machines.

Practitioner insights:

The viability of a single brain across forms hinges on a robust, safety-validated software stack and a governance framework that can handle different morphologies without introducing reliability risk.

The economics rely on deep, scalable component sourcing; cost advantages from volume depend on rapid adoption and the resilience of the core component supply chain.

Early welding deployments will test not just technical capability but repeatability and long-term maintenance across a production line.

Watch for ecosystem signals: if a dominant general-purpose automation brain proves transferable, downstream players—from sensors to integrators—may reorganize around this platform, impacting sourcing choices for global manufacturers.

In a factory floor reality where every minute counts, Xiaoyubot’s one-brain approach is a bold attempt to turn flexibility and scale into a competitive differentiator for China’s manufacturing backbone.

Sources

Xiaoyubot Founder Applies "Hit Product Logic" to Industrial General-Purpose Embodied Intelligence