Workflow
九章智算操作系统
icon
Search documents
九章云极:AI+浪潮下的中国样本,把最新的普惠智算服务带到服贸会
Qi Lu Wan Bao· 2025-09-15 04:53
Core Viewpoint - The 2025 China International Service Trade Fair showcases the innovative capabilities of China's intelligent computing industry, with a focus on the advancements made by DataCanvas, a leading company in this sector [1][3]. Group 1: Industry Developments - The intelligent computing industry in China is evolving through globalization and innovation, creating a competitive ecosystem that integrates AI with traditional industries [3]. - The integration of AI into various sectors, such as finance and healthcare, is accelerating the intelligent transformation of the global service ecosystem [3][8]. - The industry is undergoing three significant transformations: intelligentization, service-oriented, and inclusiveness, leading to a new service paradigm based on AI [3]. Group 2: Company Innovations - DataCanvas has introduced the Alaya NeW Cloud, which utilizes a serverless architecture and reinforcement learning to optimize AI infrastructure, significantly reducing the complexity of developing and applying large models [4]. - The DataCanvas intelligent computing operating system is the first in China to achieve full-stack evaluation in key areas, allowing developers to easily schedule computing resources [5]. - The new generation of intelligent agents developed by DataCanvas can execute complex tasks and improve accuracy by 54.2% through training [7]. Group 3: Market Impact - The company aims to make computing resources as accessible as utilities, thereby lowering the barriers for AI applications and transforming the supply paradigm of computing power [8]. - DataCanvas has established a standardized measurement system for computing power, allowing users to pay only for the resources they consume, which promotes efficient circulation of computing resources [8][9]. - The Aladdin computing service platform targets academic research, providing cost-effective infrastructure that enhances training efficiency by 3-4 times [9].