2026世界机器人大会 | 利元亨周俊雄董事长：扎根工艺，AI才能从工具长成伙伴

近日，2026世界机器人大会（WRC）在北京顺利举办。本届大会以“人机共生，产需共融”为主题，全方位展示机器人全产业链创新成果，生动呈现智能科技赋能实体经济、服务经济社会高质量发展的强劲动能。当前，国内机器人产业持续扩容提质，创新实力稳步攀升、应用场景持续拓展；叠加全球人工智能技术加速迭代，智能机器人已成为连通数字技术与物理实体经济的核心载体，深度赋能产业转型升级。利元亨周俊雄董事长受邀出席具身智能赋能装备与工业技术升级主题活动、参与高端制造业论坛对话，围绕行业四大核心议题分享制造业AI创新实践与前沿研判，并见证《AI链化制造》战略研究报告正式发布。

工艺适配难、落地难，是当前工业AI规模化应用的核心壁垒。工业生产直面复杂多变的物理场景，材料特性差异、工艺参数微调、生产环境波动，都会直接影响生产成效与产品品质。脱离制造工艺积累、缺乏物理规则数字化转化能力，再先进的算法模型都无法落地实效。立足高端制造深耕经验，利元亨始终致力于打通“工艺大脑”与车间现场的深度联结，所有AI研发立项均源于一线真实生产痛点，所有算法迭代优化均经过产线全工况实战验证，以良率提升、产能增效、成本优化等实打实的生产指标，检验技术落地价值。

针对通用AI与工业专用AI的差异化发展，周俊雄董事长表示，通用大模型覆盖面广、泛化能力强，但难以适配工业精细化生产需求；而利元亨自研工业智能体，是深耕垂直工序的专业智能模型，堪称制造业的“专职工匠”，具备三大核心优势。一是深耕工艺本源，依托企业多年沉淀的数十万条工艺参数、缺陷样本、设备运行数据建模，深度绑定生产机理与物理规则；二是聚焦闭环落地，摒弃单纯问答式交互，可实时优化工艺参数、预判设备隐患，直接联动装备控制系统解决实操难题；三是工业级高可靠，所有智能判断均有工艺依据可追溯，能够适配产线7×24小时不间断高强度生产，满足严苛的工业运行标准。

在AI规模化落地路径上，周俊雄董事长指出，制造业AI升级的核心瓶颈不在于顶尖算法技术，而在于复合型人才缺口。当前行业存在明显的供需脱节：算法研发人员精通模型架构与代码开发，却不熟悉车间工艺逻辑；一线资深工艺骨干拥有数十年实操经验，却难以将积累的工艺经验转化为AI训练能力。制造业的智能迭代，从来不是实验室闭门研发的成果，而是一线工艺经验持续赋能、持续滋养的结果。

现场分享中，周俊雄董事长以企业实际案例具象化展现工业AI落地价值。依托全资子公司舜势测控的检测技术积淀，针对焊缝检测行业共性难题，利元亨创新人才应用模式，安排拥有十余年实操经验的资深焊接工艺师转型“程序训练师”。无需编写代码，工艺师专注将虚焊、凹坑、瑕疵判定等专属实操经验，拆解为AI可识别的标注规则，持续迭代优化检测模型。最终，这套融合3D点云与工艺AI的智能检测方案，检测精度达0.05毫米，是人眼识别极限的2倍，单件扫描仅需0.8秒，检测效率较人工提升十倍以上，实现产线100%全覆盖精准检测。

周俊雄董事长最后总结强调：工艺是智能化的根基，根扎进工艺里，AI才能从工具长成伙伴。

谈及 2030 年产业展望，周俊雄董事长表示，随着技术持续迭代、产业生态日趋成熟，AI 将深度融入生产与生活，成为人人可及的可靠伙伴。

立足长远产业布局与国家战略需求，利元亨依托深耕智造领域多年积淀的技术、工艺与整厂交付优势，将工业AI落地实践成果转化为产业升级动能，前瞻布局三大前沿核心赛道。在AI算力领域，深耕GPU服务器精密组装技术，筑牢人工智能大模型运行的硬件根基；在智能终端领域，搭建AI眼镜等新型智能终端自动化产线，助力消费级智能产品规模化、高品质量产；在具身智能与人形机器人领域，突破核心零部件量产难题，构建从关节模组制造到整机装配的全流程智造体系。未来，利元亨将持续以高端精密装备为基石，打通AI技术从研发到落地的物理链路，赋能人工智能全产业链自主可控，为我国强化战略科技力量、推进智能制造升级注入持久动力。

2026 World Robot Conference | Lyric Chairman Zhou Junxiong: Only by rooting itself in craftsmanship can AI grow from a tool into a partner.

The 2026 World Robot Conference (WRC) was successfully held in Beijing recently. Under the theme of “Human Robot Symbiosis, Industry Demand Integration”, this conference comprehensively showcased innovative achievements across the entire robot industrial chain, vividly demonstrating the robust momentum of smart technology empowering the real economy and serving high quality socioeconomic development. Against the backdrop of continuous expansion and quality improvement in China’s robotics industry, innovation capabilities are steadily growing and application scenarios keep expanding. Coupled with the accelerated iteration of global artificial intelligence technologies, intelligent robots have become a core carrier connecting digital technologies with the physical real economy, deeply empowering industrial transformation and upgrading. Mr. Zhou Junxiong, Chairman of Lyric, was invited to the themed event on Embodied Intelligence Empowering Equipment and Industrial Technology Upgrading and took part in the high?end manufacturing industry forum dialogue. He shared practical experience and forward?looking insights on AI innovation in the manufacturing sector around four core industry topics, and witnessed the official release of the Strategic Research Report on AI?Chained Manufacturing.

Chairman Zhou Junxiong’s Speech: Rooted in Craftsmanship, Forging Intelligent Industrial Partners

At the forum, Chairman Zhou Junxiong delivered a keynote speech centered on the core theme of "AI Empowering Factory Operations". He put forward a key judgment: for industrial AI to move beyond a mere tool and evolve into a genuine symbiotic partner for the manufacturing industry, it must be deeply rooted in frontline workshop scenarios and underlying production craftsmanship, enabling intelligent technologies to take root and function in real production environments.

Difficulties in process adaptation and on site deployment constitute the core barriers to large scale application of industrial AI today. Industrial production faces complex and volatile physical scenarios. Variations in material properties, fine tuning of process parameters, and fluctuations in production environments all exert direct impacts on production performance and product quality. Without accumulated manufacturing process know how and the capability to digitize physical rules, even the most sophisticated algorithm models cannot deliver practical on site results. Drawing on its in depth experience in high end manufacturing, Lyric is committed to forging deep connections between the “Process Brain” and actual workshop operations. All AI R&D projects stem from real world pain points on production floors, and every algorithm iteration and optimization undergoes full condition practical verification on production lines. The value of technological deployment is measured by tangible production metrics including higher yield rates, improved productivity and optimized costs.

Regarding the differentiated development of general?purpose AI versus industrial?specific AI, Chairman Zhou Junxiong noted that general large?scale models feature broad coverage and strong generalization capability, yet struggle to adapt to the refined production requirements of industry. By contrast, Lyric’s self?developed industrial agents are professional intelligent models deeply cultivated for vertical manufacturing processes. Acting as “full?time craftsmen” for the manufacturing sector, they boast three core strengths. First, deep dive into the essence of manufacturing processes. Built upon hundreds of thousands of process parameters, defect samples and equipment operation data accumulated over years within the enterprise, the model establishes deep linkage between production mechanisms and physical rules. Second, focus on closed loop deployment. Moving beyond simple question and answer interaction, it can optimize process parameters in real time, predict potential equipment risks, and directly connect with equipment control systems to resolve practical operational challenges. Third, industrial grade high reliability. All intelligent judgements are traceable with process based evidence. It adapts to 7×24 hour non stop high intensity production on production lines and complies with stringent industrial operation standards.

On the path toward large scale AI implementation, Chairman Zhou Junxiong pointed out that the core bottleneck in manufacturing AI upgrading lies not in cutting edge algorithm technologies, but in the shortage of interdisciplinary talents. The industry currently faces a notable mismatch between supply and demand. Algorithm researchers excel in model architecture and coding development yet lack familiarity with on site process logic. Meanwhile, senior frontline process specialists boast decades of hands on experience but struggle to translate accumulated craftsmanship expertise into AI training capabilities. The intelligent iteration of the manufacturing industry is never an outcome of closed door laboratory R&D, but a result of continuous empowerment and enrichment drawn from real world frontline process experience.

During his on site sharing, Chairman Zhou Junxiong illustrated the tangible value of industrial AI deployment with real world enterprise cases. Leveraging the testing technology heritage of Shunshi Measurement & Control, its wholly owned subsidiary, Lyric addressed common industry challenges in weld inspection by innovating its talent application model: senior welding process engineers with more than ten years of hands on experience were retrained as “program trainers”. Without writing any code, these process engineers translate practical know how for judging cold solder joints, pits and defects into labeling rules recognizable by AI, enabling continuous iteration and optimization of the inspection model. The resulting intelligent inspection solution combining 3D point cloud and process driven AI achieves a detection accuracy of 0.05?mm — twice the resolution limit of the human eye. Each single part scan takes only 0.8?seconds, delivering over ten fold improvement in inspection efficiency versus manual work, and realizing full coverage 100 percent accurate inspection across production lines.

In his concluding remarks, Chairman Zhou Junxiong emphasized that craftsmanship is the bedrock of intelligentization, and that only by firmly rooting itself in craftsmanship can AI evolve from a mere tool into a true partner.

Looking Ahead to 2030: Deepening Roots in Smart Manufacturing to Empower Industrial Self-Reliance and Strength

Sharing his vision for 2030, Chairman Zhou Junxiong stated that as technology continues to evolve and the industrial ecosystem matures, AI will become deeply integrated into both production and daily life, transforming into a reliable partner that is accessible to everyone.

With a long term industrial layout aligned with national strategic priorities, Lyric leverages its accumulated strengths in technology, craftsmanship and turnkey factory delivery built up over years of deep engagement in intelligent manufacturing. It translates the practical outcomes of industrial AI deployment into momentum for industrial upgrading, and has proactively laid out three core frontier tracks. In the AI computing power segment, Lyric deepens its expertise in precision assembly of GPU servers, consolidating the hardware foundation for the operation of large scale artificial intelligence models. In the intelligent terminal segment, it builds automated production lines for new intelligent terminals such as AI glasses, supporting large scale, high quality mass production of consumer grade intelligent products. In the embodied intelligence and humanoid robot segment, it breaks through mass production bottlenecks of core components, and establishes a full process intelligent manufacturing system spanning from joint module fabrication to complete machine assembly. Going forward, Lyric will continue to take high end precision equipment as its cornerstone, bridge the physical link between AI technology R&D and on site deployment, and enable independent, controllable development across the entire artificial intelligence industrial chain, injecting sustained impetus into strengthening China’s strategic sci tech strength and advancing the upgrading of intelligent manufacturing.