At COMPUTEX 2026 in Taipei, held from June 2-5 under the theme 'AI Together,' Longsys unveiled its latest advancements in edge AI storage technology, showcasing a full suite of solutions designed to support next-generation AI computing workloads.
Alongside its premium consumer brand Lexar, the company demonstrated integrated hardware and software solutions aimed at optimizing local large language model (LLM) performance and accelerating ecosystem collaboration amid the global expansion of AI infrastructure.
Positioning its exhibition around 'Edge AI Storage • Integrated Implementation,' Longsys introduced two newly developed memory technologies — AIDIMM™ and AILPBGA™ — both specifically engineered for edge AI inference scenarios. The company emphasized that these innovations address the growing demand for higher bandwidth, lower latency, and improved energy efficiency in AI agent devices, embedded systems, and compact computing platforms.
AIDIMM™ is presented as a high-performance single-module memory solution capable of supporting stable operation of peripheral LLMs with more than 70 billion parameters. With capacity up to 128GB, a native 256-bit bus architecture, and claimed single-channel bandwidth up to 307.2GB/s, the module is designed to reduce computing bottlenecks in AI workloads. Longsys highlights its compact design, simplified PCB integration, and compatibility with major AI host motherboards, enabling easier system upgrades without significant hardware redesign. The module also includes dynamic voltage regulation and proprietary power management technology to optimize energy efficiency, thermal regulation, and sustained performance under continuous AI computing and multitasking.
In addition to AIDIMM™, the company introduced AILPBGA™, a compact high-bandwidth memory chip specifically designed for space-constrained embedded AI applications. With a 22×22mm form factor, customizable capacity from 24GB to 64GB, and a proprietary 256-bit bus with bandwidth up to 307GB/s, it is intended for light to medium-scale edge LLM deployments. Longsys positions the chip as a cost-effective and power-optimized alternative for embedded systems, offering broad LPDDR compatibility while significantly increasing performance density compared to traditional LPDDR5x solutions. Its architecture is designed to shorten development cycles and reduce integration complexity for device manufacturers.
Overall, Longsys framed its COMPUTEX 2026 presentation as part of a broader push to advance edge AI computing infrastructure, where storage and memory systems are becoming increasingly critical for enabling real-time intelligence, local inference, and scalable AI agent applications.