自变量再获数亿元新融资：今年已宣布融资具身智能公司中估值最高

Core Insights - The company, Zivariable Robotics, has recently secured several hundred million yuan in new financing, led by major industry capital including SAIC Capital and CICC SAIC Fund, with participation from existing investors such as Meituan, Sequoia China, and ByteDance [2] - Following a previous financing round of 1 billion yuan announced in January, Zivariable has become the highest-valued company in the field of embodied intelligence for 2026 [2] - Zivariable is the only embodied intelligence company in China to receive investments from Meituan, Alibaba, and ByteDance simultaneously [2] Company Overview - Founded in December 2023, Zivariable focuses on developing a self-researched "General Embodied Intelligence Model," utilizing real-world data to create versatile robots capable of fine operations [4] - The company's WALL-A model integrates a novel VLA architecture with world models, achieving embodied multimodal thinking chains and enhancing the robot's ability to perform tasks in unstructured environments [4][5] - Zivariable has pioneered large-scale real machine data collection and developed various data acquisition devices, ensuring a continuous supply of high-quality data through a model-driven data pipeline [5] Technological Advancements - The evolution of Zivariable's foundational model has unlocked the potential for high-degree-of-freedom dexterous hands, enabling robots to master human-like skills such as tool usage and precise finger control [5] - The company has open-sourced its end-to-end embodied foundational model, WALL-OSS, in September 2025, promoting the open dissemination of embodied intelligence technology [5] Market Applications - Zivariable has begun to penetrate high-value sectors such as industrial manufacturing, logistics, and elder care, demonstrating its robots' ability to meet real market demands with high generalization and low deployment costs [7] - With the involvement of industry capital like SAIC, collaborative efforts will enhance the deployment of embodied intelligence in automotive scenarios, further expanding the model's autonomous adaptation and operational capabilities in complex environments [7]