©Trending AI Editorial Team

RedBear AI, a company specializing in AI memory science and technology, has recently declared the successful closure of a Series A+ funding round, securing hundreds of millions of RMB, with its post-investment valuation nearing RMB 3 billion.

This latest funding round was co-led by Zhejiang Jiuwei Private Equity Fund and Jiaxing Zhangyuan Venture Capital, with participation from existing shareholder Geruifeng.

This achievement marks RedBear AI's sixth funding round in a mere 15 months. The capital raised will primarily fuel foundational brain-inspired research in memory science, facilitate the deep integration of the OpenBear general-purpose large model with the MemoryBear memory system, and expand market presence across four key application domains: intelligent customer service, intelligent marketing, ChatBI, and intelligent education.

RedBear AI's Funding History. Image Source: IT Juzi

Established in April 2024, RedBear AI was founded by Wen Deliang, who previously served as a marketing R&D engineer at Alibaba's Taobao and held technical management roles at renowned enterprises such as Fosun Group and Panshi Group. The core technical team comprises individuals who have been involved in the R&D of large models like QianWen and Zhipu AI, amassing extensive experience in large model training and inference infrastructure.

RedBear AI's foray into the realm of 'memory science' was driven by the team's firsthand encounters with the challenges of large model deployment—currently, nearly all large models exhibit a structural shortcoming of 'instant forgetting,' rendering them incapable of reusing experience in long-term tasks. Furthermore, the team posits that memory capability is the linchpin for AI's transition from the digital to the physical world. In real-world settings such as industry, logistics, and manufacturing, AI must not only 'comprehend the present' but also 'recall the past and anticipate the future.'

Presently, RedBear AI has developed the 'Super Memory Storage Framework' based on the MemoryBear brain-inspired memory engine, digitizing the episodic memory mechanism of the human hippocampus and enabling a comprehensive memory lifecycle encompassing perception, extraction, association, and forgetting. The system accommodates multimodal inputs, including text, images, voice, and sensor data, empowering AI to not only 'acquire knowledge' but also 'retain experiences and habits.'

MemoryBear Human-Like Brain Memory System

Building upon this foundation, RedBear AI has further expanded into the CodeBear memory-based programming large model platform. Unlike conventional AI programming tools, CodeBear continuously accumulates project experience, coding standards, and business logic from each interaction, mitigating the issues of contextual discontinuity and repetitive communication expenses.

In terms of commercialization, the AI customer service product has attained a 98.4% self-service resolution rate, with confirmed revenue reaching RMB 135 million for the full year of 2025. Ernst & Young has audited the pure software gross margin to range between 60% and 78%.

RedBear AI is set to officially unveil the memory-driven general-purpose large model OpenBear for consumer-facing applications on July 31, concurrent with the launch of the CodeBear programming assistant.

Founder Wen Deliang disclosed that the company anticipates submitting its listing application to the Hong Kong Stock Exchange in the second or third quarter of 2027 and officially commencing its IPO in 2028.