CATL Leads 1.1 Billion RMB Financing for Galaxy General, Setting New Record for Single Largest Embodied AI Funding in China This Year

June 23, Galbot announced that it has completed a new round of financing totaling 1.1 billion yuan, led by CATL (Contemporary Amperex Technology Co., Limited). This round attracted top-tier investors including CATL's listed company strategic investment arm, Puquan Capital (CATL Capital), China Development Bank Sci-Tech Innovation Fund, Beijing Robotics Industry Fund, GGV Capital, and others.

As the leading industrial strategic investor, CATL will continue to deepen industrial synergies, providing key support for the technological implementation and large-scale application of Galbot's embodied AI large models in the industrial sector.

It is understood that over the past two years, Galbot has raised more than 2.4 billion yuan in cumulative financing.

Public information shows that in June 2024, Galbot received a 700 million yuan angel-round financing from strategic and industrial investors such as Meituan Strategic Investment, BAIC Industrial Investment, SenseTime Guoxiang Fund, and iFlytek Fund, as well as financial institutions including Qiming Venture Partners, Blue Sky Consulting, Matrix Partners China, Source Code Capital, and IDG Capital. Lightstone Capital served as the exclusive financial advisor for this financing and participated in early-stage investments.

In November 2024, Galbot completed another 500 million yuan strategic round of financing. Investors included SAIC Hengxu, Hong Kong Investment Corporation (HKIC), Shanghai Artificial Intelligence Industry Fund, Shenzhen Capital Group (SCGC), Beijing Robotics Industry Fund, CCB International, Zhiyou Scientist Fund, Rongyi Investment, and Jincheng Capital. Existing shareholders such as IDG, Matrix Partners China, Blue Sky Consulting, and Beijing Artificial Intelligence Industry Fund also significantly increased their investments.

In June 2023, just one month after its establishment, Galbot completed its seed-round financing with investors including Matrix Partners China and Blue Sky Consulting.

On the product front, Galbot focuses on the research and development of general-purpose robots based on embodied multimodal large models. In June 2024, it released its first-generation embodied large model robot, Galbot (G1). Subsequently, Galbot launched end-to-end large models for grasping, retail, and navigation: GraspVLA, GroceryVLA, and TrackVLA.

In January this year, Galbot (Yinhe Tongyong), in collaboration with the Beijing Academy of Artificial Intelligence (BAAI) and researchers from Peking University and the University of Hong Kong, officially released GraspVLA, the first comprehensive end-to-end embodied grasping foundation model capable of broad generalization. This model achieved a global first: zero-shot (Zero-Shot) generalization capability relying solely on pre-training.

The end-to-end model GroceryVLA, designed for retail commercialization, addresses the challenges of intelligent grasping and scene adaptation in complex retail environments.

In scenarios featuring densely stacked shelves with a wide variety of goods, the model can precisely grasp items across all categories—including soft bags, hard boxes, bottles, glass containers, and plastic packaging—without requiring separate parameter tuning for each product type.

Furthermore, the model possesses strong autonomous decision-making capabilities (such as dynamically selecting optimal operation targets) and excellent anti-interference abilities (capable of handling unexpected shifts or toppled cargo). Crucially, it is easy to deploy, requiring zero scene pre-collection, which significantly reduces deployment time.

Beyond operational models, Galbot also launched TrackVLA, a product-grade end-to-end navigation foundation model. This model achieves zero-shot generalization in complex scenarios through pure visual environmental perception and natural language instruction understanding.

Additionally, it features high-dynamic target tracking capabilities. The model can accurately track human or animal targets amidst dense crowds, supports dynamic target switching, and can quickly reposition lost targets based on spatial intelligence and large model reasoning. It also demonstrates high environmental robustness, maintaining stable and reliable tracking performance in unfamiliar environments like unsurveyed shopping malls, under varying lighting conditions, or amidst obstacle interference.

In terms of scaling up robotic applications, Galbot has made significant progress in scenarios such as smart retail, industrial settings, and healthcare/elderly care.

In March this year, Galbot released the world's first humanoid robot smart retail solution. The wheeled dual-arm robot Galbot executes a fully automated process including inventory checks, restocking, picking and delivery, and packaging for 5,000 product categories, 6,000 aisles, and over 10,000 items in unmanned stores of 50 square meters around the clock. Deploying a new store requires only one day. Currently, nearly ten stores in Beijing have been deployed and are operating normally, with the potential to be put into use in hundreds of stores nationwide by the end of the year.

Meanwhile, Galbot is also continuously expanding its presence in the automotive sector. Just last week (June 17), Galbot announced the establishment of a joint venture, "BoYin HeChuang," jointly with Bosch Group's Bosch Capital, and signed a tripartite strategic memorandum with Bosch China and Bosch Capital.

The joint venture will focus on high-precision manufacturing scenarios such as complex assembly, promoting the large-scale implementation of embodied AI in the industrial sector. At the same time, the joint venture will actively promote financing expansion and global layout, creating an internationally competitive intelligent manufacturing product system.

According to the introduction, this cooperation coincides with a critical juncture where global manufacturing is accelerating towards intelligence, marking an important milestone in the transition of embodied intelligence technology from verification experiments to industrial implementation.