Lenovo Unveils "AI Factory" Solution with Industry's First Training-Inference Service Validation

On December 5, Lenovo Group hosted the "Heterogeneous Intelligence, Local Engine" 2025 Lenovo Heterogeneous Computing Industry Alliance Summit and AI Computing Infrastructure New Product Launch in Beijing.

At the event, Chen Zhenkuan, Vice President of Lenovo Group and General Manager of China Infrastructure Business Group, officially introduced the "Lenovo AI Factory" solution. The company also launched its next-generation high-end AI server for large model training—the Lenovo Wutian WA8080a G5, powered by Intel® Xeon® 6 processors—along with the upgraded Lenovo Wanquan Heterogeneous Computing Platform 4.0.

During the summit, Lenovo and industry partners jointly released the validation results of the industry's first standard for model training-inference services, titled "Information Technology—Computing Power Services—High-Performance Training-Inference Service Capability Requirements." Lenovo served as the lead editor, participating in the entire standard-setting process.

The Lenovo Wanquan Heterogeneous Computing Platform 4.0 delivers significant performance improvements: - In large-scale model pre-training scenarios, training time is reduced by 35%. - In post-training scenarios, optimizations in computing engines, few-shot reinforcement learning, and automatic parallel training cut training time by 50%. - For inference tasks, ROCE network load balancing enhances bandwidth utilization by 60%, communication primitive performance by 30%, and large model inference performance by 30%. - In hyper-intelligent fusion computing, the platform supports both international and domestic hardware ecosystems, optimizing from foundational architecture to operators. It also adds over 16 manufacturing application templates and scripts for efficient hyper-intelligent integration.

Additionally, the Lenovo Wutian WA8080a G5, Lenovo's first high-end AI server for large model training, features Intel® Xeon® 6 processors and supports the latest OAM2.0 GPUs for localized deployment. It offers dual cooling solutions—air and liquid—for enhanced thermal management.