Skip to main content
Back to News
Product
2 min read

Wanxun Technology (万勋科技) launches NOVA2.0 embodied-AI brain, claims 10 million real-world robot task executions

The AMW Read

Second-gen product launch with concrete deployment scale (10M+ task executions, 40+ industries) advances the single-brain/multi-form-factor thesis for the segment, but doesn't resolve an open debate or introduce a new top-tier entrant.
NoveltySignificance
Robotics · Player MapRobotics · Structural Forces

Wanxun Technology (万勋科技) launches NOVA2.0 embodied-AI brain, claims 10 million real-world robot task executions

Wanxun Technology released NOVA2.0, its second-generation "flexible embodied brain" architecture, describing it as the first open-world, all-weather, commercial-delivery-grade embodied-AI system in the sector. The company says the architecture has already powered more than 10 million real-world task executions across over 40 industry scenarios, including construction, power and energy, transportation, autonomous driving, and manufacturing. NOVA2.0 pairs with Wanxun's Pliabot soft-robot bodies and can drive multiple form factors — muscle-joint units, soft single- and dual-arm systems, and dexterous hands or humanoid platforms — from one shared brain. The company cites 20-millisecond trajectory generation, a claimed ability to generalize to new scenarios using roughly 1% of conventional training data, and operating tolerances from -40°C to 60°C and altitudes above 5,000 meters.

The launch reflects embodied AI's shift from lab demos and pilot proof-of-concepts toward volume commercial deployment in unstructured, high-variance environments. Wanxun's pitch — one brain generalizing across many physical form factors, improved by a data flywheel drawn from live commercial operations rather than simulation — is a direct bet that field data diversity, not model scale alone, is the binding constraint on robot generalization in the open world.

For builders and investors evaluating physical-AI plays, Wanxun's disclosed traction — over 50 large enterprise customers, deployment across more than 100 countries, and repeat orders exceeding 40% of commercial order value — is a self-reported commercial signal, not an audited benchmark. The variable worth tracking is whether cross-form-factor data reuse actually compounds into faster generalization than rivals training on curated or simulated datasets, since that is the core defensibility claim behind the "one brain, many bodies" architecture.

#EmbodiedAI #Robotics #PhysicalAI #Wanxun #RoboticsHardware

#Wanxun Technology#NOVA2.0#embodied AI#Pliabot#robotics

How This Connects

Based on Robotics · Player Map

  1. 1d agoOpenAI hires actuator and motor engineers to design its own robot hardware.OpenAI
  2. 1w agoWanxun Technology (万勋科技) launches NOVA2.0 embodied-AI brain, claims 10 million real-world robot task executions · THIS ARTICLE
  3. 1w agoPaXini opens Shenzhen showcase detailing tactile-to-execution pipeline for physical AIPaXini
  4. 2w agoTashi Zhihang's AWE3.7 Model Shows Cross-Scenario Generalization From Factory Floor to HomeTashi Zhihang

More news from Wanxun Technology

Stay updated with the latest news and announcements from Wanxun Technology.

View all Wanxun Technology news

Discover AI Startups

Explore 5,000+ AI companies with VC-grade analysis, funding data, and investment insights.

Explore Dashboard