PhysBrain 1.0
Search documents
独家 | 3个月融3轮,2026“物理AI”黑马诞生
投中网· 2026-03-27 02:30
Core Insights - The article discusses the emergence of DeepCybo, a company in the embodied intelligence sector, which has attracted over 60 investment institutions in just three weeks, indicating strong market interest and potential for growth [4][8]. - DeepCybo has successfully raised several rounds of financing, totaling several hundred million yuan, positioning itself as a key player in the embodied intelligence market for 2026 [4][8]. - The company has launched its new model, PhysBrain 1.0, which utilizes a "human-first perspective" data approach and has achieved state-of-the-art results in international evaluations [4][8]. Company Overview - DeepCybo is the first embodied intelligence enterprise incubated by Beijing Zhongguancun Academy and Zhongguancun Artificial Intelligence Research Institute, focusing on a unique "human-first perspective" technology route [4][8]. - The company has built a comprehensive data processing pipeline, accumulating approximately 300,000 hours of human real data and 100,000 hours of high-quality multimodal data in less than three months [8][12]. Technology and Innovation - The "human-first perspective" approach is not only a strategic direction for DeepCybo but also aligns with NVIDIA's recent focus on embodied intelligence, addressing industry challenges such as data scarcity and generalization [6][9]. - DeepCybo's PhysBrain 1.0 model demonstrates remarkable flexibility and adaptability, allowing robots to learn and adjust strategies in real-time, showcasing a significant advancement over traditional methods [11][18]. Team and Expertise - The founding team of DeepCybo consists of top experts in large models and practical embodied intelligence, including Chen Kai, who has extensive experience in AI and has led significant projects in the field [12][13]. - The team includes a diverse group of PhD graduates from leading institutions, contributing to a robust research and development environment that supports the company's innovative goals [17][18]. Market Position and Future Goals - DeepCybo aims to create a Chinese version of Generalist AI, focusing on building a general base model that is tailored to local industries, recognizing the importance of high-quality annotated data [15][16]. - The company is positioned to leverage its early advancements in the "human-first perspective" approach to gain a competitive edge in the rapidly evolving field of embodied intelligence [9][15].