On July 20, Juchip Technology announced that it has made significant progress in the R&D of its second-generation in-memory computing technology. Compared to the first-generation in-memory computing NPU, the second-generation technology greatly enhances computing power, supports larger model scales, and optimizes energy efficiency, utilization, and quantization accuracy across the board. SoC chips equipped with the second-generation in-memory computing NPU are set to enter formal tape-out this year, with mass adoption across multiple brands and categories expected next year. Additionally, the company is heavily investing in AI development tool platforms to build an ecosystem that helps customers efficiently adapt AI models. As the company's R&D efforts advance, the application scenarios for its edge AI chips will gradually expand beyond audio markets, continuously driving edge AI from a technological concept to a real consumer market within a diverse ecosystem of multiple brands, categories, and algorithms. (Jinshi)
All Comments