QingCloud's QingCloud Intelligence Computing division, in collaboration with TCL, has established a unified AI infrastructure foundation for the entire group. This project has been recognized as a benchmark case for implementing AI technology in industrial settings, earning the accolade of "2025 China Big Data Industry Annual Leading AI Infrastructure Enterprise." The selection was jointly organized by the Golden Ape Committee, Data Ape, and the Shanghai Big Data Alliance. Amid the global acceleration of AI technology industrialization and the wave of intelligent transformation in manufacturing, TCL Group, a global leader in consumer electronics, faced multiple challenges including fragmented computing resources, difficulties in cross-regional operations and maintenance, and heterogeneous training environments. Addressing TCL's strategic need for "AI in real-world applications," QingCloud Intelligence Computing leveraged its full-stack technical capabilities honed over a decade of enterprise-level service. They built a collaborative implementation system across six dimensions—resource investment, system architecture, core technology, product support, innovative thinking, and team coordination—to create a tailored, end-to-end solution covering both the resource and operational layers.

QingCloud Technology, in partnership with the Dongyang Sunshine Group, co-developed a "New Manufacturing Intelligent Computing Base," which was successfully recognized as a "2025 Digital Transformation Excellence Case in Smart Manufacturing China." This award was presented by the 15th Annual 2025 Smart Manufacturing China Business Leadership Selection, hosted by Digital Business Age, following a rigorous multi-stage selection process including online nominations, expert review, and online voting. Leveraging the QingCloud AI Intelligent Computing Platform, Dongyang Sunshine established a digital transformation system featuring unified computing resource scheduling and management, deep scenario adaptation, and intelligent, efficient operations and maintenance. After the project's implementation, Dongyang Sunshine achieved a leap in R&D efficiency: the design cycle for new drug molecules was reduced from 18 months to 12 months, a 33% efficiency increase; simulation time for high-performance composite material microstructures decreased by 40%; cross-facility model collaboration efficiency improved by 60%; and无效 waiting time for R&D personnel was reduced by 87.5%. In terms of cost control, through computing optimization and process upgrades, Dongyang Sunshine saved approximately 400 million yuan annually in experimental costs and reduced operations management costs by over 30%, achieving a dual breakthrough in "efficiency enhancement and cost optimization."

QingCloud officially released CloudESight v6.0, redefining the IT infrastructure platform with its golden combination of "AI Infrastructure 3.0 Base + Virtualization Core." This launch represents a dual revolution integrating technological reconstruction and ecosystem upgrade, achieving a milestone leap characterized by "100% channel-oriented, industry standardization, and extreme ease of use." CloudESight v6.0 completely breaks away from the constraints of traditional IaaS architecture, fully transitioning to a cloud-native technology stack based on "K8s + Containerd + KubeVirt + Kube-OVN + KubeSphere." This not only overcomes the limitations of poor scalability and weak ecosystem compatibility in traditional architectures, enabling seamless integration with the cloud-native ecosystem, but also deeply optimizes the SPDK vhost-user solution. It independently implements two core capabilities: "high-performance logical volume global pre-allocation" and "concurrent safe allocation of large-page memory." Featuring a plugin-based design, it supports independent upgrades and rollbacks, fully complying with K8s standards. By focusing on three core dimensions—code optimization, deployment efficiency, and operational assurance—it comprehensively reduces enterprise IT operational pressure and accelerates business implementation.

The core challenges for K8s are no longer limited to basic deployment and maintenance but are gradually shifting to three key dimensions: standardized multi-cluster governance, lean resource management, and standardized management of heterogeneous infrastructure. To help enterprises address these challenges, KubeSphere officially launched version v4.2.1, aiming to build a stable, efficient, intelligent, and cost-effective next-generation cloud-native infrastructure platform. The main updates and upgrades in KubeSphere v4.2.1 include targeted improvements in areas such as smooth gateway upgrades, multi-cluster governance, and fine-grained node scheduling. For multi-cluster scenarios, platform capabilities have been continuously optimized in terms of upgrade management and status synchronization. A new Node Group capability allows physical or virtual nodes to be logically divided into multiple groups, supporting the binding of node groups to enterprise workspaces. It provides a flexible and scalable K8s cluster inspection framework through KubeEye. By integrating Vertical Pod Autoscaler (VPA) and event-driven autoscaling mechanisms, while enhancing traditional Horizontal Pod Autoscaler (HPA) policies, it achieves more precise, agile, and multi-dimensional resource elastic scheduling capabilities. Focusing on unified management of heterogeneous infrastructure and data access efficiency, it introduces three core foundational capabilities for general heterogeneous computing scenarios like engineering simulation and industrial digital twins, providing stable, standardized computing support for upper-layer scheduling platforms.

QingCloud officially launched an independent deployment tool—QCE-ImageFlow—redefining the image management and resource migration experience for QingCloud Enterprise Cloud with a core focus on "decoupled dependencies, comprehensive functionality, high efficiency, and security." Enterprise Cloud customers, regardless of their version, can easily access more professional image and resource management services, further boosting operational efficiency. The key breakthrough of QCE-ImageFlow is its complete decoupling from the IaaS platform. It does not depend on a specific Enterprise Cloud version, requires no upgrade of the main product, and enables full functionality upon independent deployment. Its underlying logic is based on Enterprise Cloud's stable image management architecture, causing no interference to the existing cloud platform, truly achieving "out-of-the-box" usability. This allows users on different Enterprise Cloud versions to simultaneously benefit from the latest image management capabilities. It possesses six core capabilities: one-stop management of multi-format images, reducing costs and errors while improving efficiency; flexible and controllable image export, solving cross-platform migration challenges; multi-source file transfer, eliminating "local waiting"; offline virtual machine migration for seamless cross-cloud resource flow; asynchronous task processing with real-time progress visualization and easy troubleshooting; and direct deployment on image servers, enhancing both efficiency and security.

The results of the fifth series of selections for the XinChuang (Information Technology Application Innovation) awards, jointly organized by XinChuang Zongheng and the Disaster Recovery Backup Alliance XinChuang Working Committee, with co-organization by several authoritative platforms including XinChuang Consulting and The Path to XinChuang, have been officially announced. QingCloud Technology, ranking fifth in total votes, was successfully selected as one of the "2025 Top Ten XinChuang Cloud Brands" (the total vote count combined public votes and votes from certified members of XinChuang Zongheng). The "Foresight 2026" annual list selection has concluded, with QingCloud Intelligence Computing successfully making the Enterprise List's "New Buds TOP30." This list, created by the International Venture Capital Agglomeration Zone and Xiaofanzhuo, adheres to core principles of "objectivity, fairness, and professionalism." It covers enterprise lists, investment lists, and various sub-sectors including artificial intelligence, embodied AI, and advanced manufacturing, aiming to identify benchmark companies with genuine technological breakthroughs and industrial empowerment capabilities. The Enterprise List's "New Buds TOP30" specifically focuses on innovative enterprises in a high-growth phase, possessing technological moats and scalability potential. The organizers firmly believe: "True innovators are long-termists who maintain strategic focus amidst uncertainty, adhere to technological faith amidst浮躁, and continue to create incremental value for the industry even in challenging times." This successful selection signifies the organizers' comprehensive affirmation of QingCloud Intelligence Computing's technical strength, commercial value, and commitment to long-termism.