On August 12, Cambricon conducted its semi-annual performance briefing for 2026. During the session, Chairman and General Manager Chen Tianshi responded to market anxieties surrounding computing power requirements, supply chain inventory management, and the progress of new product R&D. He emphasized that as large model technologies rapidly evolve, the demand for intelligent computing is on a steady rise. Cambricon remains committed to driving technological advancements in artificial intelligence chip design, bolstering its core strengths, actively expanding its market presence, and expediting the deployment of real-world applications.

In the internet industry, Cambricon will hone in on key applications like large models, multimodal systems, and search recommendations. The company plans to refine its product technology by focusing on computing performance, communication efficiency, and framework capabilities, thereby enhancing the overall effectiveness of its offerings in internet-based environments.

Chen Tianshi also disclosed that, in the first half of 2026, Cambricon is developing a new-generation intelligent processor microarchitecture and instruction set. These developments are geared towards optimizing large model training and inference scenarios, with the goal of boosting competitiveness in areas such as programming flexibility, user-friendliness, performance, power efficiency, and chip area utilization.