SenseTime Spin-off Targets Factory Floors with TX01

A SenseTime offshoot bets wheeled humanoids will run the factory line.

Shenzhen Tianyuanxing Technology Co., Ltd. (Tianyuanxing) has closed an angel round worth tens of millions of yuan, led by Hongde Investment, Shengshi Hongyuan, and Haiyuan Capital. The company—born from SenseTime’s industrial robotics unit in December 2024 and now a SenseTime X-Ecosystem member—centers on a wheeled-chassis humanoid platform designed for industrial tasks. Founder Kuang Zhanghui, a PhD from HKU who previously ran SenseTime’s robotics unit and helped deploy lithium-battery perception systems at CATL, is steering the effort into on-floor automation.

Its core product, TX01, blends a dual‑arm humanoid silhouette with a wheeled base and a telescopic electric-cylinder system to tackle material handling, assembly, and inspection. The design signals a push to reduce line-changeover friction in factories that still rely on semi-robotic handling for heavy throughput. TX01’s specified capabilities include a 100kg payload and about four to eight hours of operation per charge, with an integrated control architecture that the team says ties perception, planning, and manipulation into one stack. Prototypes were completed in 2025, with mass deployment targeted for 2026.

The customer list underscores a bridge between AI-driven perception and traditional manufacturing corridors. CATL and Sunwoda, two major battery-pack and cell manufacturers, appear among early clients, alongside CREC Electrification Bureau Group, a state-linked contractor with extensive rail and infrastructure projects. The alignment with CREC and similar heavy-assembly players hints at a path for humanoid robots to tackle not just automated palletizing but more complex assembly and inspection tasks in electrified propulsion and energy storage lines. The company’s connection to SenseTime’s ecosystem also signals a continuity of AI-vision know-how feeding robotic manipulation on the shop floor.

From a Chinese industrial robotics lens, Tianyuanxing’s move is telling. It’s rare to see a startup launch with the cultural and technical leverage of a SenseTime lineage, then accelerate capital-raising at a “tens of millions” scale in a single round. The arrangement suggests both strong private funding appetite for factory-automation bets and a belief that a well‑positioned perception stack—perceived by many as SenseTime’s core strength—can be married to practical, field-ready actuators and control flows. The TX01’s claimed 2026 mass deployment window places it squarely in China’s ongoing push to upgrade manufacturing with smarter automation while balancing cost, reliability, and integration with existing MES and SCADA ecosystems.

For global manufacturers watching China’s robotification wave, the Tianyuanxing story crystallizes a few near-term truths. First, the integration hurdle remains real but is narrowing when you can buy a solution with a single, end-to-end control architecture rather than stitching disparate perception, motion, and robot-control layers. Second, tier-one clients in batteries and infrastructure-heavy industries signal demand for robots that can withstand heavy handling and sustained uptime across multi-shift lines. Third, the ecosystem advantage matters: a SenseTime-backed team with access to a broad AI-vision network may compress the time-to-value for factories seeking faster deployment of automated lines.

Two practitioner-level takeaways stand out. One, the TX01’s 100kg payload and eight-hour ceiling imply a sweet spot for mid-to-heavy material handling and assembly tasks where rigid gantries may be less economical. Manufacturers should map where a humanoid platform truly wins over fixed robots or gantries—areas with frequent reconfiguration or human-robot collaboration where vision-driven picking, inspection, and placement add value. Two, the onus will be on system integrators to deliver seamless MES/ERP interfacing; the “integrated control architecture” promises less bespoke integration but still requires robust OPC UA/MES hooks, cybersecurity, and maintenance protocols to realize ROI in real-world lines. Finally, the risk profile is notable: a spin-off with a high-powered AI heritage must prove its hardware reliability and long-term component availability to avoid a value trap in the critical wear-and-tear components that powers every factory shift.

In a landscape where industrial robots frequently hinge on a single customer’s appetite and a narrow tech stack, Tianyuanxing’s TX01 represents a strategic bet that perception-enabled humanoids can move beyond pilot lines to steady-state production in China’s manufacturing heartlands. If 2026 brings scalable deployments, the lesson will be less about novelty and more about what it takes to blend AI-driven sight with dependable manipulation on the factory floor.

Sources

Ex-SenseTime Industrial Robotics Team Raises Tens of Millions for Wheeled Robots