1月16日，工业富联2026年度科技创新峰会在深圳举行。大会以"AI Makes AI"为主题，聚焦人工智能与智能制造深度融合，围绕产业技术演进、应用落地 与生态协同等核心议题展开深入探讨，共同展望新一轮产业升级的机遇与挑战。 本次峰会汇聚三十余位AI与智能制造领域的重量级发言嘉宾。来自产业界、学术界及投资界的近五百位嘉宾出席会议。 工业富联董事长郑弘孟在致辞中表示，随着模型越来越大，AI训练功耗几乎每年都在翻倍，算力密度也越来越高，单位空间内的能耗和散热挑战日益凸 显，其背后牵动的已不只是单一的产品，而是整个产业链的全面升级。 "未来的AI与智能制造，不是一家公司在比拼，而是一整个产业体系在共同进化。也正因如此，今天产业竞争的本质也正在发生根本的变化：从谁的技术 最好，逐渐要走向谁能把系统整合得更好。"郑弘孟表示，技术可以很前沿，但合作始终要以人为本；创新可以很复杂，但产业的进步永远来自于彼此的 连接。他期待所有与会者能用更开放的态度，一起来探索技术的边界；用更务实的方式，把创新真正落地；用更长远的眼光，打造可持续的产业生态。 主论坛 AI赋能未来：机遇、路径与挑战 【主旨演讲】构建真正好用的企业级智能体 ...

Core Insights - The "EVOLVE 2025" summit showcased the roadmap for enterprise-level AI agents and introduced a "3+2+2" product matrix to facilitate rapid development of AI agents for businesses [1] - The summit emphasized the collaboration among major cloud service providers to create a sustainable AI ecosystem through the "Super Connection" global partner program [1] Group 1: AI Development Trends - Piero Scaruffi highlighted a clear trend of technological integration in generative AI by 2025, with innovations like diffusion Transformers and multi-modal capabilities becoming standard [3] - The emergence of new technologies such as thinking chains and expert mixtures is reshaping the landscape of AI applications [3] Group 2: Evolution of AI Agents - The distinction between traditional AI products and advanced AI agents was made, with the latter being likened to autonomous driving, capable of executing complex workflows independently [4] - The operational mechanism of these AI agents is summarized as a cycle of perception, decision-making, action, and learning, allowing them to adapt to various environmental changes [4] Group 3: Multi-Agent Systems - The transition from applications to multi-agent systems introduces challenges in orchestration, necessitating a new technology stack that includes hardware, cloud services, and orchestration layers [5] - The concept of "context engineering" is emphasized, requiring AI agents to understand organizational structures and goals beyond executing single tasks [5] Group 4: Industry Applications - Various sectors are witnessing innovative applications of AI, particularly in customer support, where intelligent systems can understand context and emotions, enhancing user experience [6] - Companies like Johnson Controls have developed integrated AI systems that significantly improve efficiency in maintenance and troubleshooting [6] Group 5: Trust in AI - The "Waymo effect" illustrates the growing trust in AI as autonomous vehicles become more prevalent, laying a foundation for broader AI agent applications [7] - Scaruffi envisions a future where multiple AI agents collaborate dynamically, akin to human social interactions, to achieve common goals [7]

Core Insights - The "EVOLVE2025" summit highlighted the launch of a comprehensive enterprise-level intelligent agent roadmap by Zhongguancun KJ, featuring a "3+2+2" product matrix that includes three foundational platforms and two application platforms, aimed at accelerating the large-scale implementation of intelligent agents in various industries [1][2] Group 1: Intelligent Agent Development - The development of large models is rooted in the accumulation of smaller models and data modeling, emphasizing the need for data to be transformed into knowledge through the discovery of hidden patterns [1][2] - Intelligent agents integrate core capabilities such as perception, understanding, decision-making, and control, serving as key vehicles for technology implementation [1][2] - The evolution of intelligent agents is supported by foundational algorithms like deep learning and reinforcement learning, with a focus on enhancing efficiency through collaborative deployment across cloud, edge, and endpoint [1][2] Group 2: Industry Trends and Challenges - The need for precision and lightweight models in large model deployment is critical, with techniques like model distillation helping to reduce computational requirements [2] - There are technical risks such as "hallucinations" in natural language understanding, particularly in accurately grasping Chinese semantics, which remain a long-term challenge [2] - The future direction involves transitioning large models and intelligent agents from general-purpose to specialized applications tailored to specific industries and product scenarios [2] Group 3: AI Agent as a Central Hub - AI intelligent agents are seen as the central brain for enterprises, addressing issues like data silos and process fragmentation by connecting key elements such as people, resources, and systems [3] - Each connection made by intelligent agents generates new interaction data, which in turn iterates the model itself, leading to increased intelligence and value creation for enterprises [3] - The evolution from the internet to mobile internet and now to artificial intelligence represents an evolution of connectivity, with intelligent agents acting as super connectors within and outside organizations [2][3]

Group 1 - The demand for overseas AI reasoning computing power is expected to grow significantly, with projected increases in total computing power demand for AI Agent applications reaching 8 times, 3.5 times, and 2.5 times from 2026 to 2028 respectively [1] - NVIDIA's upcoming AI chips, Vera Rubin NVL144 and Rubin Ultra NVL576, are expected to outperform previous models significantly, with performance increases of 3.3 times and 14 times respectively compared to the GB300 NVL72 [1] - The supply-demand gap for overseas reasoning computing power is anticipated to widen due to the rapid growth in demand outpacing the evolution of AI chip performance [1] Group 2 - The overall monthly active user growth for domestic AI applications is lagging behind that of overseas applications, with the top 20 overseas AI products showing a monthly growth rate of approximately 4% as of May this year [2] - Doubao has demonstrated a relatively strong performance, with its daily token usage exceeding 16.4 trillion, while Huolanganjing holds a 46.4% market share in the domestic public cloud for large model calls [2] - The positive cycle between overseas AI computing power and applications is expected to drive significant capital expenditure guidance from major North American cloud providers by 2025 [2] Group 3 - The business model of AI Agents is shifting from "providing tools" to "delivering value," indicating a potential revaluation opportunity for SAAS companies [3] - The enterprise-level AI Agents are at a critical point of scalable application, with priority sectors for implementation identified as enterprise services (OA/ERP/CRM), finance/risk control, marketing/e-commerce, manufacturing/supply chain, legal/government, and healthcare/education [3] - The growth of productivity AI agents is being driven by efficiency improvements, with a notable increase in AI tools for programming, education, and creative sectors [3]