On June 25th, Lenovo held the "Together, Reaching New Heights: Lenovo Tianwen Brand Refresh and Computing Power Ecosystem Conference."

At the event, the Lenovo Group Tianwen brand officially launched its refreshed strategic vision and introduced three core achievements: the Wanquan Heterogeneous Intelligent Computing Platform V5.0, a hyper-node solution, and the industry monograph "Token Factory."

Lenovo Group Vice President and General Manager of the China Infrastructure Business Group, Chen Zhenkuan, announced ambitious growth targets: "By 2027, Lenovo's China Infrastructure Group aims to anchor a goal of 100 billion RMB and is targeting the number one position in China's server market."

Regarding the brand strategy, Lenovo Tianwen has completed a transition from a "localized server brand" to a "leader in China's AI computing power infrastructure." This shift is oriented towards enterprise AI solutions, driven by "AI + R&D innovation," and grounded in AI computing power assurance, systematically reconstructing technology, product, and ecosystem capabilities for the AI era.

Furthermore, at this conference, Lenovo also released the Wanquan Heterogeneous Intelligent Computing Platform V5.0 and the hyper-node solution.

This platform features two core breakthroughs: its cluster training and inference acceleration technology, utilizing core technologies like hierarchical decoupling of the PD separation architecture and KV Cache shared cache optimization, significantly enhances cluster resource utilization. Its chip-model compilation optimization technology enables adaptive matching of computational graphs for different models and automatic generation of operators.

Concurrently, Lenovo also premiered the Lenovo Tianwen Hyper-Node Computing Solution. This solution boasts four key advantages: Supercomputing power, with a single node capable of hosting 40 GPUs, delivering over 28 PFLOPS of FP8 compute power, and breaking 5.76 TB of HBM memory, meeting the extreme demands of training and inference for trillion-parameter large models; Fully interconnected low latency, with total memory access bandwidth exceeding 80 TB/s and chip-to-chip P2P one-way latency at the hundred-nanosecond level; Flexible scalability, supporting 40-card single nodes and scale-out horizontal cluster expansion, while also being backward compatible with 32-card configurations; Simplified deployment, utilizing a cable-less orthogonal direct-insertion architecture compatible with standard 19-inch racks, compressing cluster deployment cycles to just a few hours.

Additionally, the conference saw the debut of the official Lenovo Group industry monograph "Token Factory: The Wealth Production Paradigm of the Token Economy." The book, centered around the four core concepts of "Token Factory," "Intelligent Production Capacity," "Effective Computing Power," and "Security and Controllability," comprehensively deconstructs the underlying logic of how infrastructure in the AI era evolves from computing power support to an intelligent production system.