人工智能处理器NPU

Search documents
直击业绩说明会 | 瑞芯微:公司将推出更高算力芯片,助力AI大模型端侧落地
Mei Ri Jing Ji Xin Wen· 2025-04-30 12:12
Core Viewpoint - The rapid evolution of AI models and their deployment in edge computing is expected to drive significant growth in the AIoT market over the next 5 to 10 years, with applications across various industries such as automotive, robotics, education, healthcare, and agriculture [1][2]. Group 1: Company Insights - Ruixinwei highlighted the transformative impact of AI models, particularly the advancements made during the recent Spring Festival, which have led to improved performance and usability of open-source models [1]. - The company noted that the development of edge AI applications is being accelerated due to the decreasing barriers to entry for AI technology, which benefits from local processing advantages such as real-time capabilities, low network dependency, and enhanced privacy protection [1]. - Ruixinwei has established a comprehensive layout of AIoT SoCs with built-in NPU capabilities ranging from 0.2 TOPs to 6 TOPs, catering to various market needs and computational requirements [2]. Group 2: Future Developments - The company plans to continue enhancing its NPU technology to support larger-scale edge models, with a focus on performance improvements and higher computational power in future chip releases [2]. - The next-generation flagship chip, RK3688, is expected to provide not only higher performance but also greater flexibility for deploying larger-scale edge models through multi-chip cascading and co-processor expansion [2].