存储器跃升AI运行关键基础设施！这家A股公司上半年净利同比增长715倍
Memory Emerges as Critical Infrastructure for AI Operations! The A-share Company Sees Its H1 Net Profit Surge 715-fold Year-on-Year
【环球网财经综合报道】8月10日晚，在披露上半年业绩的A股上市公司中，最受市场关注的是存储龙头江波龙。2026年上半年，该公司实现营收240.88亿元，同比增长136.26%；实现归母净利润105.77亿元，同比增长71528.66%。
[Global Times Finance Report] On the evening of August 10, among A-share listed companies releasing their first-half results, Longsys, the storage sector leader, drew the most market attention. In H1 2026, the company posted revenue of 24.088 billion yuan, a year-on-year increase of 136.26%; its net profit attributable to parent company stood at 10.577 billion yuan, representing a year-on-year surge of 71,528.66%.
进一步来讲，江波龙主要业务为存储器和主控芯片的研发设计、封装测试、技术支持与存储器销售。该公司通过覆盖芯片设计（主控芯片及存储芯片）、固件算法开发、存储器设计、封装测试等存储关键环节的全栈式技术能力，为市场提供消费级、企业级、车规级、工规级存储器以及行业存储软硬件应用解决方案。
Furthermore, Longsys' core businesses cover the R&D, design, packaging and testing, technical support and sales of memory devices and controller chips. The company provides the market with consumer‑grade, enterprise‑grade, automotive‑grade and industrial‑grade memory devices, as well as industry‑specific storage software and hardware application solutions through its full‑stack technical capabilities covering key storage links such as chip design (controller chips and memory chips), firmware algorithm development, memory‑device design, and packaging and testing.
半年报中，江波龙表示，自2025 年第三季度以来，在 AI相关应用及数据中心基础设施建设对存储的庞大需求支撑下，半导体存储市场进入快速扩张期。在云端市场上，AI 模型的训练与推理伴随着密集的数据读写，拉动了AI 服务器及数据中心对大容量、高带宽及高可靠性存储器产品的需求。在端侧市场上，AI手机、AI电脑、智能汽车、AI可穿戴设备及具身智能设备的推出，使本地数据生成、模型执行及实时交互不断增加，对存储器的性能、功耗及尺寸提出了更为严苛的新要求。在此背景下，存储器已不再仅仅是被动的数据载体，而是逐渐演化为AI运行的关键基础设施。
In the semi-annual report, Longsys stated that since the third quarter of 2025, driven by the massive demand for memory from AI-related applications and data center infrastructure construction, the semiconductor memory market has entered a period of rapid expansion. In the cloud market, the training and inference of AI models involve intensive data read/write operations, driving the demand for large-capacity, high-bandwidth, and highly reliable memory products in AI servers and data centers. In the edge market, the launch of AI smartphones, AI PCs, smart vehicles, AI wearables, and embodied AI devices has led to a continuous increase in local data generation, model execution, and real-time interaction, imposing more stringent new requirements on memory performance, power consumption, and form factor. Against this backdrop, memory is no longer merely a passive data carrier but has gradually evolved into critical infrastructure for AI operations.
同日晚间，江波龙发布关于回购公司股份方案的公告。公司拟使用自有资金或自筹资金以集中竞价交易方式回购公司股份，回购资金总额不低于4亿元且不超过8亿元，回购价格不超过735元/股。回购股份将用于股权激励或员工持股计划，本次回购股份的实施期限为自公司董事会审议通过回购方案之日起12个月。（南木）
Later that evening, Longsys released an announcement regarding its share repurchase plan. The company intends to use its own funds or self-raised funds to repurchase shares through centralized bidding trading, with the total repurchase amount of no less than RMB 400 million and no more than RMB 800 million, at a repurchase price not exceeding RMB 735 per share. The repurchased shares will be used for equity incentive plans or employee stock ownership plans. The implementation period for this share repurchase is 12 months from the date on which the repurchase plan is reviewed and approved by the Board of Directors of the company. (Nan Mu)
