At the inaugural Open Compute Project (OCP) APAC Summit in Taipei, ASE Group Executive Vice President Tingu Yin Chang said surging AI computing demand has become the primary catalyst for innovation across the semiconductor sector. Advanced packaging technologies now face three critical challenges: increasing package size, escalating power consumption, and complex thermal management.
Samsung Electronics has secured an order from Tesla for artificial intelligence (AI) chips, offering a renewed boost to its foundry business. However, the company continues to face significant challenges in the memory sector, particularly in the high-bandwidth memory (HBM) segment, where it has yet to pass Nvidia’s quality tests, which could hinder its performance recovery.
Huawei plans to open-source its Compute Architecture for Neural Networks (CANN), a heterogeneous AI computing framework positioned as a direct alternative to Nvidia’s dominant CUDA platform.