LUISUAN Technology's "Defining a New Level of AI Inference Storage: G3.5 Level KV Cache Solution" was selected as an outstanding achievement in AI intelligent computing innovation at the 2026 Smart Expo
National platform authoritative certification, domestic AI storage independent innovation sets a new benchmark!
On September 11, the 16th Smart City and Smart Economy Expo (hereinafter referred to as the "Smart Expo") opened grandly at the Ningbo International Convention and Exhibition Center. During the exhibition, the organizing committee held a special press conference on outstanding achievements in AI intelligent computing innovation and officially announced the selection results. After self-application by the unit, preliminary review by the organizing committee and comprehensive review by industry experts, and rigorous evaluation based on multiple dimensions such as technological innovation, advanced computing power, practicality, industry demonstration and generalizability, LUISUAN Technology’s “Defining a New Level of AI Inference Storage: G3.5 Level KV Cache Solution” stood out from many participating results across the country and was officially selected as an outstanding achievement in AI smart computing innovation at this Smart Expo.



This year's Smart Expo, with the theme of "AI smart change, smart future", is co-sponsored by the Ningbo Municipal People's Government, Zhejiang Provincial Department of Economy and Information Technology, China Academy of Information and Communications Technology, China Electronics and Information Industry Federation, China Telecom, China Mobile, and China Unicom. It is a highly influential national professional platform in the field of domestic smart economy and smart cities. Being able to win this award on this platform fully demonstrates the technological leadership and industry leadership of LUISUAN Technology in the field of AI inference storage.
Cracking the "Video Memory Wall" and Defining a New Level of AI Inference Storage
Currently, large AI models are accelerating their evolution towards long-context and high-concurrency reasoning. In order to alleviate the high cost and waste of resources caused by insufficient single-card video memory (HBM) capacity in traditional tightly coupled architectures, the G3.5 storage layer built on the LUISUAN Technology GP7000 platform realizes the decoupling of GPU computing power and KV Cache storage by pooling KV Cache into a bare block-level shared flash memory pool, which is becoming an important evolution direction of large-scale, long-context AI inference architecture.
From technological breakthroughs to industrial demonstrations, let AI infrastructure "spend smarter"
This selection is not only a high recognition of the innovative value of LUISUAN Technology by industry authorities, but also a sign that domestic AI storage solutions have occupied a place in the evolution of the world's most cutting-edge AI reasoning architecture. From overcoming the "video memory wall" to defining the ultimate architecture of "storage is memory", LUISUAN Technology is working with industry chain partners to promote the intelligent computing center from integrating storage and computing to the era of minimalist and efficient storage and computing separated computing power pools, contributing professional strength to the implementation of "artificial intelligence +" actions and the cultivation of new quality productivity.