本财季，联想ISG基础设施方案业务集团营收367亿人民币，同比增长超31%。其中，云基础设施(CSP) 业务持续受益于AI训练需求扩张，延续高景气度；企业基础设施(E/SMB)业务营收增长的同时，通过定 价管理优化与运营效率提升，实现毛利率持续改善。区域层面，ISG中国业务实现关键突破，企业基础 设施成为利润的重要贡献来源。 值得关注的是，本财季联想ISG业务亏损继续大幅收窄。公司介绍称，这主要得益于ISG于第三财季启 动战略性重组。预计该重组将在未来三年带来每年14亿人民币的净成本节省，推动盈利能力持续改善。 值得注意的是，本财季AI服务器收入保持高速增长，并拥有高达155亿美元的订单储备。同时，联想披 露了其部署了首个基于GB 300 NVL72的机架式解决方案，并同步完善产品矩阵，发布包括AI推理服务 器SR675i、SR650i和边缘计算服务器SE455i在内的推理优化产品组合，覆盖从数据中心到边缘侧的多层 级推理应用场景。此外，受益于企业级推理需求持续释放，其海神液冷技术业务同比增长300%，进一 步巩固该公司在高能效AI基础设施领域的市场地位。 2026年2月12日，全球AI生态链领先企业联想集团 ...

Core Insights - Lenovo has moved its global innovation technology conference to CES 2026, marking the largest scale in the event's history, with key figures from major tech companies like NVIDIA, AMD, Intel, and Qualcomm in attendance [1] - Lenovo's CEO emphasized that the new wave of computing power will stem from the explosion of AI inference [1] Group 1: AI Strategy and Product Development - Lenovo is focusing on "hybrid AI," which encompasses diverse AI models to meet user needs and the ability for cross-device collaboration among smart devices [3] - The company introduced three key technologies: Intelligent Model Orchestration, Agent Core, and Multi-agent Collaboration, aimed at matching models to user needs, understanding user intent, and enabling collaborative AI operations [3] - Lenovo's AI super intelligent agent, Lenovo Qira, was launched for overseas markets, showcasing cross-device collaboration capabilities among various devices [5] Group 2: Cloud and Local Model Integration - The Qira model has shifted from purely local calls to a combination of cloud and edge collaboration, although local models may lose about 40% of their capabilities compared to cloud models [5][6] - Lenovo's AI strategy involves optimizing partner models rather than developing them in-house, with a focus on integrating cloud models into their existing scheduling system [6] - The company plans to implement cross-device task distribution, allowing users to execute commands across devices [6] Group 3: Partnerships and Market Positioning - Major chip manufacturers are aligning with the trend of enterprise AI transformation, with NVIDIA's CEO stating that companies need to integrate advanced models with their data to build proprietary intelligent systems [8] - Lenovo announced a significant partnership with NVIDIA to establish an "AI Cloud Super Factory," aiming to support large-scale AI models and quadruple their business collaboration over the next 3-4 years [9] - Lenovo will also be among the first to adopt AMD's Helios architecture and collaborate with Intel on AI PCs, indicating a strong commitment to AI-driven enterprise solutions [10]