Huawei Unveils Fifth OlympusMons Award Winners, Advancing AI Storage Innovation

[Shenzhen, China, September 22, 2025] At the 2025 Global Data Storage Expert Forum, the winners of the fifth OlympusMons Awards were announced, consisting of six leading global research teams that stood out for their breakthroughs in data storage. The OlympusMons Award was presented to Professor Wu Yongwei and his team from Tsinghua University, which developed the innovative solution—Trading More Storage for Less Computation—to address the performance and efficiency bottlenecks in large model inference. The OlympusMons Pioneer Award was awarded to five other teams.

Dr. Peter Zhou, Vice President of Huawei and President of Data Storage Product Line, presents the OlympusMons award to Professor Wu Yongwei and his team from Tsinghua University

As a leading competition in the data storage industry, the fifth OlympusMons Awards focus on addressing critical technical pain points in the AI era, with particular emphasis on two directions: Storage Technologies That Deliver Ultimate Per-Bit Cost Efficiency, and New Data Foundations for the AI Era. This initiative centers on tackling issues such as limited bandwidth between xPUs and storage, low utilization of computing clusters, extended inference latency, and rising data storage costs driven by rapid data growth and increased data value.

The awards attracted 95 experts and scholars from 19 universities and research institutions worldwide. From this competitive field, a rigorous review process selected just six winning teams representing the absolute best in data storage research and innovation. The team led by Tsinghua’s Professor Wu won the highest honor, the OlympusMons Award, for its technical achievement of Trading More Storage for Less Computation: High-Performance LLM Inference System. The other five research teams won the OlympusMons Pioneer Award with cutting-edge technical achievements, including:

• Approximate Retrieval Technology of High-Dimensional Vector for Multimodal Data by the Zhang Kai team from Fudan University
• AI Data Foundation Driven by Efficient Cache Management and Multimodal Large Models by the Zhou Ke team from Huazhong University of Science and Technology and Peking University
• High-Density Magnetic and Optical Storage Media Technology with Superior Cost-Effectiveness by the Zhang Jingyu team from Huazhong University of Science and Technology
• Hyper-Compression: Model Compression via Hyperfunction by the Fan Fenglei team from City University of Hong Kong
• Declarative Data Processing on Massive Heterogeneous and Distributed Data Sets and Streams by the Volker Markl team from Technische Universität Berlin

At the forum, Chang Sheng, director of the R&D Management Department of Huawei Data Storage Product Line, announced the two directions and four challenges of the sixth OlympusMons Awards. The first is Novel Media Technologies for the AI Era, which includes two key challenges: high-performance, high-capacity AI SSD technology and warm data storage media technology. The first direction aims to break through the bottlenecks of existing storage media technologies and meet the urgent demand of AI workloads for high-performance and large-capacity storage media. The other direction is Agentic AI–Native Data Foundation, which includes two key challenges: multimodal data management and knowledge generation technology, and data acceleration networks and architectures for AI. This aims to explore core high-performance data network technology and build an intelligent data platform for multimodal data processing and knowledge generation.

The OlympusMons Awards were set up by Huawei in 2019 to encourage global academicians to work on data storage research and core technologies that can be rolled out in the industry, thereby promoting mutually beneficial outcomes for the industry, universities, and research institutes.